CONTINUATION--GLOSSARY OF TERMS AND TECHNIQUES FOR POLITICAL ANALYSIS

• Document Type: CREST
• Collection: General CIA Records
• Document Number (FOIA) /ESDN (CREST): CIA-RDP85T00875R001100020002-2
• Release Decision: RIFPUB
• Original Classification: K
• Document Page Count: 37
• Document Creation Date: November 9, 2016
• Document Release Date: February 18, 1999
• Sequence Number: 2
• Publication Date: January 1, 1974
• Content Type: SUMMARY

 App roves ~,`r'',tffayg~ilk, si '61 vd t0s`T~ Wpaance c,s
49
MEAN - 61
65
I
-1.90 MEAN ?1.1.96 Z SCORES
0
zero-sum game-a particular class of problems dealt with by game
theory. The zero-sum game is a special case of the fixed-sum game, which
describes purely competitive situations in which one player's gains are equal
to the other player's losses. No cooperation between players is possihle in
these games where the sum of the payoffs to the two players in each cell in
the matrix is set at some constant value. It is set at zero in the zero sum
case, but any value will do in general.) (see: GAME THEORY,,fIxed-sum
game, saddle point)
Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
 Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
TECHNIQUES
Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
 Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
ACTION TI] ORV-a body of thought identified with
structural-functional analysis, in which the basic units of analysis are the
social and political actions of individuals. Action in the political sense
usually relates to the efforts of political decision-makers in making and
implementing public policy, or to the behavior of'states in the international
arena. Analysis of the patterns of action can lead to an understanding of
political decisions and, in some ca:;es, provide a basis for prediction. (see:
STR UCTURA L-FUNCTIONA L ANAL YSIS, political action)
ANALYSIS OF VARIANCI (ANOVA)-a procedure fir making a
statistical inference about the source of two or more sample. groups. The
intent is to compare the mean values of the samples in order to discover
whether or not the samples were drawn from the same overall population.
The results, read from an F table, then tell how often the observed
variability in the sample means could be expected to result from mere
random fluctuation, and how confident we may be in the conclusion that
at least one of the observed samples was drawn from a different population.
Analysis of variance can be used, for example, in analyzing voting results.
If we know the average vote for a particular party's candidates in each of
several regions, ANOVA will c.ll how likeiy it is that some special factors
other than just chance variations have uniquely influenced the voting in
one of the regions, without significantly affecting the others. This alerts us
to any special characteristics of the unusual area, and further investigation
can be directed toward it. (see: F test)
BAYESIAN ANALYSIS-a formal method for using incoming data
to modify previously estimated probabilities. Each new piece of informa-
tion may be evaluated and combined with prior historical or subjective
assessments of the probability of an event in order to determine whether its
occurrence has now been made more or less likely, and by how much. Baye-
sian analysis can also be used to compute the likelihood that the observed
data are attributable to particular causes. One advantage claimed for Baye-
sian analysis is its ability to blend the subjective probability judgments of
experts with historical frequencies and the latest sample evidence.
Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
 Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
BEHAVIORALISM (not to be confused with behaviorism)--an orienta-
tion toward the study of social phenomena which is characterized by em-
piricism, quantitative data manipulation, and 'applied research. This is
distin';t from the traditionalist attitude, which relies more heavily upon
historicism, introspection, and intuition. Within the last 30 years
behavioralism has received widespread acceptance, especially in psy-
chology, economics, and political science. It is partially an attempt to make
the social sciences more "scientific" (i.e., more like the natural sciences) with
a concern for rigorous research methodology, empirical ,ibservation, and
numerical information. (see: behaviorism)
BUREAUCRATIC POLITICS-a generic term for a theoretical ap-
proach to the study of the interplay between individuals within
organizations or governments. The approach focuses primarily upon the
goals, motivations, and incentives of important individuals in the
bureaucratic hierarchy and attempts to predict the outcomes of their
decisions which are often less influenced by the substantive issues at stake
than by personal preferences and infighting.
In his book, Essence of Decision, Graham Allison uses the bureaucratic
politics model (which he calls Model II I 'l to analyze the decisions made dur-
ing the Cuban missile crisis of 1962 by the US and USSR. He visualizes the
US and Soviet governments not as unitary actors or as aggregations of
organizations, but as arenas of competition between individuals. Allison
proposes that the process of decision-making through bureaucratic politics
is analogous; to the process through which the international system comes to
be what it is at any given moment-a balanc% among the various powers
that make it up. (see: rational actor model, ORGANIZATION
THEORY)
CASE STUDY-research aimed at uncovering detailed information
about a particular political event or phenomenon. The case study deals with
a single observation or event, as contrasted with comparative studies or the
sampling of many cases through some form of survey research. Case studies
are, however, often used to generate hypotheses and to test hypotheses as
they apply to a specific event. (see: SUR VEY RESEARCH, hypothesis)
Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
 Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
CLUSTER ANALYSIS-a technique for the orderly classification of
objects or individuals into small groups on the basis of shared
characteristics. A large, unorganized collection of items with many
dif'f'erent attributes is divided into smaller similar subgroups. Cluster
analysis clumps together individuals who are as alike as possible, and as
different from those in the other groups as possible. There is an element of
arbitrariness as to precisely where to draw group boundaries-they may
overlap or they may be forced apart. Some cluster techniques split
hierarchically with smaller groups nested inside broader ones.
Cluster analysis differs from DISCRIMINATION ANALYSIS in that the
former discovers groups while the latter begins with recognized groups and
assigns new individuals to them. An example of a hierarchical cluster
analysis would be an attempt to determine the likely coalitions and factions
within the ruling elite in a given country. By examining family ties, jointly
held political ideologies, common economic interests, etc., a cluster analysis
can predict who the likely partners are, and how they will combine with
others to form larger united groups on various issues.
The Rice-Beyle cluster-bloc technique is a method for processing votes in
such a way as to reveal clusters of agreeing voters. Once the issues to be
analyzed are selected, the frequency with which every voter agrees with
every other voter, and a matrix of agreement is constructed. The highest
agreement scores are than selected and "clustered" together. The computa-
tions art laborious and usually done by computer.
COHORT ANALYSIS-studies which examine the impact of political
variables upon members of different age groups or other time identified
groups. Suppose an analyst wants to compare the nature of political at-
titudes between two generations (say, under 40 and over 40) in a given
government. Using cohort analysis he can measure the two groups at the
same time. Alternatively, he could measure the younger group once, and
again 20 years later, to ascertain changes in political attitudes with in-
creasing age. (see: PANEL STUDY)
COMMUNICATIONS THEORY-a body of thought that seeks to
explain processes of information sending, receiving, storing, and utilization
as they relate to a political system. Communication is the means by which a
political system receives and acts upon demands and develops policies. In-
formation concerning the effectiveness of the system's policies and decisions
Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
 Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
on the environment is termed fi.'edback, and the extent to which the system
is able to adjust to feedback information helps to determine its effectiveness
as a political system. Coin inunications theory also relates to the ways in
which attitudes and images are built up in the minds of decision makers and
the way they affect perception, values, goals, and decision-making. A com-
munications analysis of' the relations between two governments, for exam-
ple, might study the ebb and flow of communications of all types (letters,
governmental statements, official visits, private travel, economic transac-
tions, etc.) between the two.
COMPARATIVE POLITICS-the determination of differences and
similarities between two or more units of analysis. A political event or case
can be compared with other similar cases, with hypothetical cases, or with
cases occurring in a time sequence. Contemporary applications of com-
parative politics include studies of specific political events, analysis of local,
national, and cross-national institutions and systems, and comparisons of
political behavior on all levels. CROSS-CULTURAL ANALYSIS, by
contrast, compares different political cultures--or patterns of beliefs,
feelings, and orientations toward government and politics within
societies-whi'e CONTEXTUAL ANALYSIS studies the interaction
between a single political system and its environment.
COMPUTER CONIrERENCINO-a general term for a variety of
techniques for communicating via computers. Often circumstances make
actual meetings or face-to-face discussions impracticable or unproductive:
time may be urgent, expenses high, psychological inhibitions may intervene,
etc. Computer teletypes can provide immediate, convenient, anonymous in-
teraction free from many of these constraints. This mode of conferencing
also has an advantage over actual meetings in terms of convenience. The
pa. ticipants can use the computer terminals at any time without rearranging
their schedules-they do not need to coordinate with everyone else for one
set time period. Additionally, if the conference is not to be held in strict
anonymity, the computer allows the participants the ability to send
messages to certain selected individuals-a kind of "whispering" which does
not often occur in actual group meetings. These directed and private com-
munications permit the formulation of joint positions between individual
participants before presentation to the others.
Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
 Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
CONFLICT BEHAVIOR ANALYSIS-a social/psychological
method which studies aggressive human behavior to understand, describe,
and predict political behavior. Conflict behavior can be studied either as it
appears domestically as political instability, or internationally as inter-state
conflict. The Dollard-Doob (frustration-aggression) hypothesis, for exam-
ple, maintains that aggression is always a consequence of frustration which
ensues when goal-oriented behavior is thwarted. According to the Dollard
study, an extra amount of energy is mobilized whenever a barrier is inter-
posed between a person and his desired goal. This energy mobilization, if
continued and unsuccessful in overcoming the barrier, tends to overflow into
generalized destructive behavior. The degree of frustration and the number
of frustrated response-sequences determine the degree of aggression.
CONTENT ANALYSIS--a technique for systematically coding and
evaluating speeches, essays, broadcasts, etc. in order to make specific in-
ferences about their origin, author, sponsor, intended audience, or other
elements of the environment. It may simply count the number of times cer-
tain words and phrases appear or it may code the tone or meaning on - scale
like friendship/hostility. With a computer, comparisons may be made, for
example, of the reactions of two individuals on a particular issue, or of the
trend in one individual's speeches. By noting the incidence of key words and
phrases, content analysis can also determine which of several speeches or es-
says were written by the same individual. This knowledge can then be used
to gauge the relative influence of certain advisors, or to discover factions
and coalitions within the inner circle of a government.
CONTEXTUAL OR CONFIGURATIONAL ANALYSIS-study of
the interaction and relationship between a particular political system and its
environment, including other political systems. Studies which relate a
political system to its physical and social environment excluding other
political systems, are termed ECOLOGICAL ANALYSIS. (see:
ECOLOGICAL ANALYSIS, political system)
CORRELATION-a statistical technique for measuring th,e closeness
of the relationship between two variables. The coefficient of determination,
tells what percentage of the variation in one factor can be explained or
predicted solely by knowledge of the other factor. Multiple correlation is
Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
 Approved For Release 1999/09/25: CIA-RDP85T00875R001100020002-2
used to measure the relationship of one factor (the "dependent variable"----
the one we want to predict) to two or more other factors (the "in-
dependent variables"-the ones used as predictors.) We compute the coef'-
f'icient of multiple determination, lt2, to tell what percentage of the observ-
ed change in the dependent variable may be accounted for by the fact that
the independent variables change. To determine how much variation is at-
tributable to each of the independent variables sepur;ttely we compute par-
tial correlation : oefficients. These resemble the coef3cient of determination
except that they measure the degree of relationship between the dependent
and independent variables if the rest of the independent variables had been
held constant instead of fluctuating as they actually do. Correlation alone
can not be used to establish causality. Generally it tells only that two
variables are associated, that they move together or that they do not. It can
not tell which one causes the other.
Correlation, which tells how closely two variables are related, is often
used in conjunction with REGRESSION analysis, which describes the
precise form ol'the relationship.
COST BENEFIT ANALYSIS---The appraisal of anticipated expendi-
tures needed to accomplish given objectives, in relation to the value of
expected gains. Examining the ratio of benefits to costs helps decision
makers to determine which among several aiternative courses of action may
offer the opportunity for the most effective use of resources. COST EFFEC-
TIVENESS is a related form of analysis that is applied to ongoing programs
to evaluate their relative progress toward the achievement of goals sought.
CROSS-CULTURAL ANALYSIS-the comparative study of the
political cultures-or patterns of beliefs, feelings, and orientations toward
government and politics-of several societies in order to describe, explain,
and understand the differences and similarities between them. (see:
political culture)
CROSS-IMPACT ANALYSIS-a forecasting method which ex-
amines the inter-relationships between variables to determine how a change
in the probability of occurrence of one event will affect the probabilities of
future related events. For example, many events would affect and simul-
taneously be affected by the advancement of European integration.
Approved For Release 1999/09/25: CIA-RDP85T00875R001100020002-2
 Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
Once the probability of occurrence of each of these events is decided, a
cross-impact matrix is constructed which lists events by number across the
top and down the left side of the matrix. The probability of each -vent's
occurrence is written above the event number across the top. Thr proba-
bility of event #6 occurring-"further enlargement of the EEC"---is 1810.
Changes in the probabilities of other events by the occurrence of event #6
are then read from the matrix. If event #6 occurred, for example, the
probability of event /f2-"stabilization of the European money market"---
would be increased from 55% to 86%. The probability of event //7--"con-
tinued independence of French military from NATO or European defense
community"-is reduced from 53% to 38%, by an increase in European
integration. From this kind of analysis, we can predict the likihood of
conceivable future events, given the occurrence of one event.
,non.mmllte
t,tnl MUVn"B
71
1
!!
7
!7
!
ID
6
33
7
H
!
AB
9
67
10
JJ
Ii
76
17
1
!
28
53
53
07
35
00
74
76
76
6
7
71
00
30
V--
13
27
52
10
35
.
01
76
e
39
33
9 '
36
10
57
00
36
10
33
11
17
10
57
7!.
53
67
67
I~
55
57
22
53
27
DECISION-MAKING ANALYSIS-a method of analysis used to
gain an understanding of the way a political system functions through ob-
serving and analyzing its decision processes: perception of the problems,
marshaling facts and information, considering alternatives, and choosing a
course of action calculated to gain chosen objectives. The decision-making
analyst tries to answer such questions as what was the decision, who made
it, how was it made, why was it made, what impact did the environmental
situation have, and what do these answers tell us abott,t the functioning of
the political system that produced the decision. Decision-making analysis
usually assumes rational behavior on the part of decision-makers, so,
generalizations concerning political behavior can be made.
Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
 Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
DECISION THEORY-a body of knowledge and related
mathematical techniques drawn from statistics, mathematics, and logic
designed to select tl: a optimal decision or chain of decisions under con-
ditions of uncertainty, Begin by listing the opening alternatives and the
probabilities of various future conditions. Additional choices may be re-
quired by each future state. The series of alternatives branching from each
decision point is graphically shown in a decision tree. The analyst can then
assign probabilities and compute the expected payoffs from each policy.
Decision analysis encourages a wider scan of alternatives and a thorough
search for future possibilities. It is similar to GAME THEORY except that
in game theory, decisions involve an intelligent opponent player while in
decision theory, the only opponent is nature, with its related uncertainty.
Thus, in situations where the opponent can be expected to act in a fashion
designed to maximize his own expected returns, game theory is the more ap-
propriate analytical tool. If the opponent plays random or non-purposeful
strategies, decision theory is the more accurate model.
DECISION TREE-a method of calculating the probability of a series
of events by listing all possible decisions and their probabilities following
each previous decision in a series. A sample decision tree may list the courses
open to a government in reacting to a hostile incident with a foreign
nation. The first decision is whether or not to act immediately. If not, the
upper br: nch of the tree is followed and the decision must then be made
whether to discuss the event at an upcoming conference, or to ignore it all
together. If, on the other hand, the government decides that immediate ac-
tion is necessary, it follows the lower branch of the decision tree. The deci-
s.on it now faces is whether to protest publicly or in private, and each of
these options entails further choices, as suggested by the remaining un-
named branches of the decision tree.
The end of the tree indicates the probability of success of each combination
of policies in preventing a repetition of the original incident. Since the
policies also have different costs in terms of risk and political expense, the
government will not automatically opt for the combination of policies which
provides the highest likelihood of success, but calculations of this type
might influence the ultimate decisions.
Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
 Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
DISCUSS
WAIT
~(.75)
I.so)
IGNORE
? .... .
INCIDENT
(.2 )
PRIVATE
ACT
I.aol
- , _ -
(,oa)
PUDLIC
DELPHI-a technique for generating and evaluating a concensus of
expert opinion. A sequence of individual questionaires is submitted, which
may in the second and successive rounds a defense of the
respondent's position or comments on the other responses. Panelists are
presented with the results and comments from the previous interrogations so
that they have the opportunity to change their views in response to the
arguments of the other participants. During the course of the series, the
source of the various responses is never revealed to the participants. This
complete absence of face-to-face confrontation minimizes the influences of
unwanted psychological factors such as personality dominance of certain in-
dividuals and the unwillingness to state unpopular views or to disagree with
superiors or associates. On the other hand, the lack of direct interaction
also eliminates the possibility of considering the source of an option when
evaluating its validity.
DIMENSIONAL ANALYSIS-dimensional analysis is the starting
point for evaluating the logic of a thought system-it checks whether the
language used makes sense. For example, in conducting a voting study, we
need to make certain that the census areas and the voting districts are iden-
tical, otherwise our demographic statistics and the polling results will not
apply to the same areas. Moreover, some key census definitions may have
changed over the years covered by the study: the terms "white collar
worker," "literacy," and "poverty" may have different standards at different
times. Dimensional analysis also concerns problems of expression. If one
figure is written in dollars, for example, and another is expressed as a
percentage, each must be changed into common units in order to be mul-
tiplied together.
Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
 Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
DISCRIMINATION ANALYSIS-a statistical technique for defining
the best boundaries separating a number of recognized groups and assigning
individuals to them. Beginning with a number of known groups and
measurements of each individual, formulas are produced which describe the
appropriate boundary lines. A "center of gravity" for each group is deter-
mined and measurements are made of the tendency to cluster or disperse
t,
tround this center. Discrimination analysis is useful not only in confirming
or amending the original group classification scheme, but in assigning new
members to the proper groups. If we know, for example, that a particular
country has three basic kinds of fighter aircraft, and that each kind has
been observed in several widely differing prototypes, discrimination analysis
can help categorize them. Based upon measurements of relevant character-
istics, the planes will be divided into three similar groups, and new obser-
vations can then be assigned to the appropriate class. Discrimination differs
from CLUSTER ANALYSIS in that the latter discovers groups, whereas
the former begins with known groups.
DYNAMIC PROGRAMMING-a complex mathematical technique
used in the solution of problems involving decisions which are grouped into
stages which then are considered sequentially. The objective is to determine
the optimal decision at each stage, with optimal referring to the entire
problem rather than to the specific stage involved. Dynamic programming
can be applied to a variety of mathematical models. For example, dynamic
programming might be appropriate in preparing a contingency plan for a
country's optimal post-war recovery program, because many variables and
constraints would have to be included in order to determine optimal short
and long run goals. (see: DECISION THEORY)
ECLECTIC APPROACH-a mode of analysis that draws from a
variety of patterns of thought or methodologies rather than focusing on a
single, comprehensive approach. An eclectic approach might combine, for
example, the use of CASE STUDIES, the SOCIAL/PSYCHOLOGICAL
APPROACH, ELITE ANALYSIS, GROUP THEORY, and
DECISION-MAKING ANALYSIS to analyze and describe the politics of
a certain nation.
Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
 Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
I:COLOGICA1.. ANALYSIS-analysis of the relationship of' a
political system to its physical and ;social ;,urroundings, excluding other po-
litical systems. Other political systems are regarded as part ol'the "context"
rather than the "ecology" and th:; study of interaction between a political
system and its context is called CONTEXTUAL ANALYSIS. Ecological
analysis calls intention to the importance of environmental factors for the
functioning of a political system and the reciprocal impact of the system
upon its environment. For example, il'a government in a developing country
institutes a broad program of land reform, its impact on the physical,
cultural, and social setting may be extensive. The results may involve a
revolution in agricultural techniques, a developing sense of nationalism, and
a general movement away from the traditional community to the large
cities. These changes in turn may have important consequen.,i for the
maintenance of the political system. (s.--e: CONTEXTUAL ANALYSIS)
ECONOMETRICS-a branch of economics which exi,resses
economic theories in mathematical terms in order to verify them by
statistical techniques. It is concerned with empirical, quantitative
measurements, and with evaluation of the cross-impact of economic
variables upon one another. An econometric model of GNP, for example,
may include a series of 300 or more multiple regression equations. The data
is based upon the historical records and the general forms of the equations
are specified, in order to predict future values.
This description and extrapolation is not only a predictive exercise, but an
explanatory one, attempting to define relationships between economic
variables and to aid understanding of their effects. It may also suggest
policy implications. Typically, the analyst knows which variables will be
important in the equations, but not their specific values-.this is supplied by
econometric analysis. For example, an economist may know that construc-
tion spending, stock averages, and unemployment are among the important
predictor variables for GNP; this analysis tells how to weigh them in mak-
ing predictions.
ELECTION ANALYSIS-an application of REGRESSION analysis
to relationships between voting results and census figures. Beginning with
the polling statistics for several comparable regions, or for one region over a
series of elections, the attempt is made to explain the observed fluctuations
in the vote for one party or individual, and to predict future voting results.
Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
 Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
Census data can provide the independent variables: differences in average
income for the regions, unemployment, literacy, etc., which will affect the
popularity of different political figures from region to region. Multiple
regression analysis we;ghts the importance of the different variables and
describes the form of the relationship. It offers predictions based upon the
best fit of the historical data.
ELITE ANALYSIS-studies aimed at explaining the nature and role
of those groups or individuals in which decision-making power is highly con-
centrated. Some studies of American elites have indicated that a single
power elite drawn from the economic upper class provides leadership for
other elements throughout the political community. Conversely, community
power studies conducted by political scientists have tended to reject the
elitist model in favor of a pluralist model in which power is portrayed as dif-
fused among many competing interest groups. Studies of foreign elites
focus on the nature and role of the ruling group in authoritarian states, the
elites that participate in the "modernization" process in developing states,
and the various elites that function within democratic societies on the local,
provincial and national levels.
ENDS-MEANS ANALYSIS-employed as an aid in problem solving
and decision-making, ends-means analysis constitutes a systematic attempt
to relate rationally selected techniques (means) to desired goals (ends) so as
to maximize goal achievement. In each case, the means are tested and
evaluated in terms of how well they perform in attaining predetermined
ends. The application of ends-means analysis may v?ry from the wholly im-
plicit to the use of rigorous, systematic causal models. (see: DECISION
THEORY, causa' model)
ENVELOPE CURVE FORECASTING-a predictive device which
approximates general trends by extending the patterns of individual com-
ponents. Envelope curves are especially useful in long range technological
and environmental forecasting. For example, if we want to predict the
dates by which various speeds of travel wiil be attainable, envelope
curve forecasting can help consider the possibilities of various scientific
breakthroughs. We graph time and speed, and chart the increasing
velocities available from various modes of transport: horseback, trains,
cars, planes, etc. Typically, each new system adds a great deal of speed at
first, then gradually contributes less and less improvement, as its
Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
 Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
technological possibilities are exhausted. A new breakthrough then occurs,
bringing another sudden burst of improvement. Therefore, if we attempted
to predict future speeds by following the trend line of any individual device
alone, its own particular curve would almost certainly understate the ap-
propriate long term expectations. An envelope curve avoids this problem by
encompassing all the different individual curves and predicting future passi-
ble breakthroughs and trends.
EVENTS ANALYSIS-based on the theory that international
relations result from the flows of discrete events or contacts between coun-
tries and that these flows are subject to minor and major fluctuations, events
analysis keeps account of these flows in order to establish a record of inter-
national relations. There are two types of events to be tra; ked: the routine
exchanges across national boundaries such as mail flows or tourists called
transactions, and the public, political flows such as official visits or govern-
ment communiques termed interactions. (see: events data)
EVOLUTIONARY OPERATION-an experimental procedure
designed to improve an ongoing process by gradual variations in its con-
trolling variables. I f given a system or operation whose yield is a function of
several different factors, evolutionary operation tries small changes in two
or three elements to move the process toward better performance. Each in-
cremental adjustment will suggest which variable to change, its direction
and amount, until an optimal solution is reached by this empirical hunt-
ing and testing.
Experimentation of this sort is applicable to many sorts of problems: what
combination of seed, fertilizer, irrigation, etc., makes a particular crop grow
best in a certain area? What combination of give-and-take in a multiple
issue negotiation process seems to elicit the most favorable response? What
combination of army, navy, and air force arms spending provides the most
efficient allocation? In a simulation model, experts might attempt estimates
of the effects of these changes, perhaps enlisting a computer model, but in
evolutionary operation, the actual changes themselves are implemented and
tried out in the field. This provides a series of controlled experiments in
order to find the optimal solution.
Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
 Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
FACTOR ANALYSIS-a group of methods for finding a few fun-
dantental (lut,ntities (the factors) underlying many observed variables. The
variables are first collected into related groups and it single measure is
selected to represent each set. The new group of basic factors-fewer in
number than the original variables-is easier to work with, to weight, and to
summarize.
Factor analysis aims to cut out some of the duplication and overwork
resulting from too much redundant data. For example, if we wish to
categorize information on 300 political leaders in some country, it may be
unnecessarily cumbersome to record and manipulate data on 20 or more
variables-such as age, education, income, and social class-for each per-
son, A simpler method would be to select only a few of the 20 variables to
represent the restr Factor analysis may show, for example, that eight of the
variables are highly associated; they seem to measure approximately the
same thing, and any one of them may be selected to represent the rest. 11'
single variables can be found to summa.' ze other groups of' variables, the
number of' variables may be cut from the original 20 to just three or four.
Any subsequent data manipulation will have only these variables to work
with, and can dispense with the rest.
FIELD THEORY-an approach to social and political analysis that
conceives of indiviipuals and grours us acti,,;o within a "field" or "life
space"-each actor's behavior being -infected by his relationshi,,I) to the par-
ticular configuration of forces in his "field." Taking countries as objects of
analysis, for example, two forces or variables might be wealth and popula-
tion. Countries could then be located in the "field," in relation to one
another, at the points indic:a.ted by their respective wealth and population.
Shifts over time in the ',ocation of such points, occasioned by attribute
change, are indicated by vectors (represented on a chart by arrows of ap-
propriate length and direction, or in mathematical models by other sym-
bols).
The field theory approach views the world as a dynamic relationship of con-
tending and complementary forces, with the objectives of identifying,
describing, measuring, and fitting each attribute into its proper niche within
the field and determining the general directions of the trends within the
force field. In international relations, field theory has been applied to the
study of rel;tionships between state attributes and state behavior. (s';e: at-
tribute, actor, mode!)
Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
 Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
EXPERIMENTAL DESIGN AND ANALYSIS--a way of discover-
ing which combinations of'variable factors best influence an outcome by set-
ting the different factors at various levels, performing experiments, and
analyzing results. This is a trial-and-error process, designed to optimize
output or discover causality by specifying a limited number of relevant
variables and closely studying the effects of changes in them. '['his is similar
to the technique of EVOLUTIONARY OPERATION, except that A does
not necessarily focus on small, incremental changes.
Experimental design and analysis can cover a wide range of possibilities,
and clever design can single out those combinations of factors which seem
most promising, thereby cutting down on the number of experiments which
must be run. This technique can also indicate a confidence level (telling how
certain we may U e of the statistical significance of the observed data) and a
sensitivity lc-!cl (telling how much the output will vary due io relatively
small changes in the input).
EXPONENTIAL SMOOTHING-a method of producing combined
forecasts for a number of different time series by the simple blending of the
last forecast and the last actual value. This is a process of combining a
number of related elements into one, and of constantly updating the
parameters based upon recent experience. For example, in predicting the
annual defense budget of some country, accurate forecasts may have dif-
ficulty taking into account all the relevant factors. By blending in the values
of the latest actual observation (i.e., the previous year's budget) a simple
formula for modifying the prediction can be made. This method is often
just as accurate as the more exhaustive techniques such as REGRESSION
analysis. (see: parameter)
EXTRAPOLATION-a prediction of the future based upon a simple
extension of present trends. Generally the most primitive test a new
theoretical model undergoes is to determine whether the predictions it offers
are more accurate, than those of simple extrapolation. Often predictions
based on trend extrapolation-for example, that the future will be es-
sentially like the past-prove to be the most reliable forecasts possible
from the available data.
Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
 Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
FLOW CHAR'.'--an orderly representation of any process, organiza-
tion, or system in which a'; iviti(;,, are defined and their interrelationships
are illustrated in graphic form. Flow charting is applicable to analysis of
any problem, but is used primarily in the initial stages of computer pro-
gramming and modeling. Lxaoiples include a graphic representation of a
bureaucratic hierarchy and the lines of communication and control within
it; a description ol' a computer program or algorithm; and a picture of a
store's traffic pattern showing where aisles interface, where queues form, or
how traffic flows.
Flow charts are shorthand devices which contain written instructions,
symbols, and occasionally equations. They should contain only enough
information to show the activity at the given level of detail desired. Some of
the more common flow charting symbols include:
1/
O
PROCESS (Computation)
A group of program instructions that perform a processing function
of the program .
INPUT/OUTPUT
Any function of an Input/output device (making information avail-
able for processing, recording processing Information, tape
positioning, etc.).
DECISION
The decision function Is wed to document points in the program
More a branrh to alternate paths Is possible based upon variable
conditions.
PROGRAM MODIFICATION
An instruction or group of Instructions which changes the program
sequence.
TERMINAL
The beginning, end, or a point of Interruption in a program.
O CONNECTOR
An entry from, or an exit of another part of the program flowchart.
FLOW DIRECTION
The direction of processing or data flow,
I Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
 Approved For Release 1999/09/25: CIA-RDP85T00875R001100020002-2
FUNCTIONALISM-are umbrella approach covering a variety of
analytic methods which focus on the firnct/o wi of a political system (the
results or consequences of actions which satisfy the needs or demands of the
political system for its own maintenance) as a means of undcrstnnding and
describing that system. (see: STRUCTURAL-FUNCTIONA ,L ANALYSIS)
CAME THEORY-u formal mathematical treatment of competitive
and cooperative situations dealing with the individual decision process of
rational adversarici-playcrs who select the. strategy which maximizes their
return and minimizes their loss. Payoff matrices are constructed which
show the outcomes and appropriate tactics of ea.;h player. For example, the
optimum level of US defense spending depends in part upon the current
level of Soviet defense spending, and a simple game theory matrix could
show this situation in contrast to Game Theory situations-in which both
players are actively seeking to advance their own self interest-are
DECISION THEORY problems where only one of the players is consid-
ered rational (for example, an individual playing against the caprice of
nature).
The most commonly drawn preliminary distinction in game theory is that
between non-zero-sum and zero-sum games. In a zero-sum game, one
player wins what the other loses. Examples of real-life zero-sum games are
electoral races between two candidates for a single post, most military tac-
tical situations, and bids by two contractors for a single project. Non-:sero-
sum games can be either cooperative or non-cooperative. In a cooperative
game the players are permitted to communicate with each other and to
exchange information in advance concerning their intended tactics. In a
non-cooperative game, communication is not permitted, although the
choice of each player becomes obvious after the play.
Games are also referred to by the number of players-N-person
games. Games involving three or more players are much more complicated
because the number of possible interacting strategies increases at an ex-
ponential rate with the number of players. N-person games are usually
used to study coalition formation. When several players are in a game to-
gether, it becomes natural for two or more to form a coalition in which case
the remaining players are encouraged to do the same in order to maximize
their gains. :f'two coalitions emerge, forcing all players to choose one or the
other, the game in effect is reduced to a two-person zero-sum game. (see:
DECISION TYEOR Y)
Approved For Release 1999/09/25: CIA-RDP85T00875R001100020002-2
 Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
GENERATIONAL POLITICS (see: COHORT ANALYSIS)
GROUP TUEORY-uo approach that seeks to explain political
behavior primarily through the study of the nature and interaction of'social
groups, Croup theorists in political science regard the group as the primary
basis for all political activity, and therefore, for political analysis, whereas
other investigators view the group as only one of several focal points for the
study of political activity. Group theory postulates that the key activities of
politics--socialization, communication, pursuit of "interest"---are interac-
tions carried on within or between groups and, therefore, should be iden-
tified and studied in that framework.
Small group theory deals with decision-making and other interactive
processes in small primary groups usually numbering twenty persons or
fewer. Such studies include both formal and informal groups, but special
attention is given to the informal relationships that supplement formal
group procedures. (see: group)
GUTTMAN SCALE-a technique of scaling used to measure attitudes
or events that can be ranked with reference to a common characterist?c. At-
titudes or events are scaled on a continuum so that an affirmative response
on any item implies affirmative responses on all other items of lower rank.
A scale measuring political instabilty over a 6 month period, for example,
might rank items in the foilowir,g manner: (1) resignation of a cabinet of-
ficial, (2) demonstrations, (3) assassination, (4) riots or mass arrests, (`.)
coup, (6) civil war. If tht scale is valid, the highest form of instability (6)
would li!.ely be preceded by all other forms (1 through 5). On the other hand,
if a number of resignations and coups occurred but no demonstrations,
assassinations, or riots, the items obviously would not scale. The scale's
"coefficient of reproducibility" would be high in the first case and low in
the latter. (see: coefficient of reproducibility, SCALING)
INSTABILITY EVENT Jul Aug Sept Oct Nov Dec
Resignation of Cabinet official x X x x
Peaceful demonstrations x x x x
Assassination x x
Mass arrests or severe riots x x x x
Coup d'etat x
Civil War X
Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
 Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
1.11ERA11CHICAL INFERENCE--a technique for decomposing com-
plex problems into a hierarchical structure and assessing the probabilitie'. Li
each level of the structure separal,ely begi:ining with the lowest level (data)
and working through intermediate variables to the highest level-the
hypotheses. By contrast, in BA YL'SIAN ANALYSIS, experts are asked to
assess the likelihood linking data directly to hypotheses. For many
problems, however, it is unlikely that one individual has the necessary ex-
perience and expertise to relate the lower level data to the upper level
hypotheses. By decomposing the problem, it is possible for the specialized
knowledge of each expert to be applied in a logically consistent manner.
For example, the hierarchical structure formulated for the hypothetical
problem of whether or not Egypt intends to develop a nuclear weapons
capability within the next five years might aan~ain four levels. The first level
consists of the exhaustive set of hypotheses: (Hi) Egypt does intend to
develop nuclear weapons and (H2) Egypt does not intend to develop nuclear
weapons. The second level contains the data (D) and activities (A) which
directly impact on the hypotheses. The third level of the hierarchy contains
those data and indicators (I) which impact on the hypotheses through the
activities described in the second level. The fourth level, logically, contains
those data which impact on the hypotheses through indicators in the third
level and activities in the second level. Different analysts are consulted for
information and judgments at each level of the problem. (see: BAYESIAN
ANALYSIS)
Hl. EGY 'T INTENDS TO DEVELOP A NUCLLAll WEAPONS
CAPADILITY WITHIN 5 YFAIIS
H2 EGYPT DOEJ NOT INTEND TO DEVELOP A NUCLEAR
WEAPONS CAPAIIII.ITY WITHIN 5 YFAIIS
DI
ISRAEL TIwEATE.NS
EGYPT WITH USE Or
DRASTIC" WEAPONS IF
TERRORISM CONTINUES
E IZI
ENRICHMENT
PLANT
EXPANSION
E_..._ D2It
PHOTO*RECON
ADDITIONAL
COOLING TOWERS
N
L UCLEAR R"D
1'ROGHAM
03
IIh D 1'l1UOITAM
DIIIECIORS MEET
IN WEEKLONG
CONFERENCE
....-.122 ~ I NO OB FHV!")~ I. L4i
OF NUCLEAR
MATER AL
t-
__ .._. D Z21'----
10
"o INCREASE IN
USE OF
HEAVY WATER
CHANGE IN
NUCLEAR ZEST
ACTIVITY
D2zz
SALES OF
r PLUTONIUM WATEIT
TO OTHER COUNTRIES
L CANCELLED
A?
IIIGIT EXPLOSIVE
LR&D PROGIIAN
D4 It
DECREASE IN
PUDLICATIOTI ON
HIGH EXPLOSIVE
RESEARCH
INCREASED
SCIENTIFIC
ACTIVITY
Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
 Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
INPUT/OUTPUT ANALYSIS-a forecasting technique especially
useful for long range conomic forecasting, Typically, it is a balance sheet
of a national economy: tht, output o1'each sector of the economy is either an
input into another sector or a consumable. The solution to this system will
yield the level of output of each industry which is required to satisfy the total
demand for each commodity consumed,
INTEGRATION THEORY-a technique of ana y,,iis which seems to
explain and measure the amount of "community" or strong cohesiveness
among peoples within a political en!ity or between entities. The three ap-
proaches to integration theory r,re: (I) federalism, (2) functionalism, and (3)
communication. The federalists study movements toward or away from in-
ternational political union. Functionalism studies the structures and func-
tions that govern economic, political and social ties between states. The
communications approach examines international transactions in order to
explain and measure the degree of cohesiveness between or within states. A
relatively new fourth approach, neofunctionalism, focuses on decision-
making and the unintended integrative consequences of previous
cooperative acts. (see: STRUCTURAL-FUNCTIONAL ANALYSIS)
INTERACTION ANALYSIS-the study of activities involving two or
more individuals or groups in which the actions of each party serve as
stimuli for responses from the other. Such stimulus-response actions may
be continuous, constituting a pattern of interact.on. Interaction patterns
between states are important to the understandii.g of inter and intra state
relations. Typical political interaction patterns involve demands, promises,
threats, bargains, coercion, and the responses to these. (see: transactions,
EVENTS ANALYSIS)
LEGISLATIVE ROLL-CALL ANALYSIS-a broad approach which
uses various techniques of quantitative analysis, such as GUTTMAN
SCALING and FACTOR ANALYSIS, to understand, explain, and
predict the voting behavior of legislators. Quantitative methods of roll-call
analysis describe and measure variations in the voting behavior of in-
dividual legislators or groups of legislators and variations among roll calls.
Coalitions, parties, blocs, and groups can be identified. Issues of domestic
and foreign policy can be compared and contrasted in terms of proponents
and opponents, and "clusters" of votes can be discovered that indicate
otherwise hidden attitudes and governmental actions.
Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
 Approved For Release 1999/09/25: CIA-RDP85T00875R001100020002-2
LINEAR PROGRAMMING-u mathematical problcm-solving
technique which selects an optimal solution, subject to one or more con-
straints. Linear programming is especially useful for determining the op-
timal mix of resources needed to achieve a desired result under certain
limitations, when the number of variables involved prohibits either
graphical methods or simple algebraic solutions. Three ingredients are
needed: (1) linear variables-that is, ones whose formulas arc described by
straight lines; (2) constraints, relations among variables, usually expressed
in the form of inequalities, which restrict the values certain variiubles may
take; (3) objective function-the rule for maximizing or minimizing which is
the target of the analysis, For example, linear programming might be used
to help a large company determine how many men, machines, and how
much money to allocate to the production of each of its various products. In
this case, the linear variables would describe the relationship between, say,
labor input and production or between capital and output; the constraints
would include the budget and the availability of needed machinery; and the
objective function would be profit or market share. Linear programming is
the simplest kind of mathematical programming, which is a general name
for several mathematical problem solving techniques. "Program inin,g" here
is not used in the sense of programming a computer, although computers
are often used in solving prob;ems of linear programming. "Programming"
here is used in the sense of "planning" for an optimal solution. (see:
MATHEMATICAL PROGRAMMING)
LONGITUDINAL ANALYSIS (see: TIME SERIES ANALKFIS)
MARKOV ANALYSIS (Markov cha!n)-a sophisticated mathematical
technique for predicting the eventual state of affairs in a system, from a
knowledge of its current situation and of the probabilities of elementary
transitions through a sequence of intermediary states. It does not give
peri,)d-by-period forecasts, bc.;t conies right to a description of the _dtimate
point the system will reach rter passing from st,~tc to state. When events
are known to be sequential and orderly, a probability diagram or a que' ing
simulation can be used to trace the probabilities of each alternative path.
Markov analysis is used when uncertainty exists about the order of events or
when the system may include feedback. For example, the behavior of con-
sumers in switching back and forth betwecr various brands of gasoline
might be the subject of a Markov analysis. It would estimate consumer
loyalty and fickleness to determine the overall anticipated market shares
of the different producers.
Approved For Release 1999/09/25: CIA-RDP85T00875R001100020002-2
 Approved For Release 1999/09/25: CIA-RDP85T00875R001100020002-2
MATIIEMATICAL PROGRAMMING-a gcnct c term for several
mathematical problem-solving techniques involving resou'ec: allocations.
The most common method is LINEAR PROGRAM' rNG, in which the
variables do not take exponential values (i,e., they may be graphed by
straight lines). More esoteric and complex programming variations include
non-linear, dynamic, stochastic, parametric, integer, quadratic, and convex
programming, each of which applies to snccialized cases and objectives.
MONTE CARLO SIMULA'T'ION--a statistical technique for syn-
thetic generation of sample data when actual samples arc unavailable.
Useful whet., experimentation would be difficult, expensive, or time con-
suming, Monte Carlo methods generate data by using random numbers and
probabilities. It thus provides a way to test a model or simulation with sam-
ple values. For example, instead of attempting to predict in advance the
probability of'di.fferent volunteer rates for the armed services, and how these
would affect manpower aliocation requirements, we could make some
assumptions about, these rates by using a Monte Carlo simulation. This
would supply rai;aom numbers, as a check on how well the system could
handle these h ypothecical values.
MORPHOLOGICAL ANALYSIS-a technique for long range
prediction, involving identification of all possible combinations of variables
which may comprise the future environment. Illogical or highly unlikely
combinations are weeded out, leaving a series of combinations or scenarios
describing possible futures. Normally, this e3ercise is an organizing tool to
develop a starting point for further analysis by other methods, especially
probability networks. Morphological analysis itself is especially relevant to
long range technological prediction of scientific breakthroughs based upon
previous discoveries. It also has applications in socio-political and
economic forecasting. At its simplest level, morphological analysis is mere-
ly organized brainstorming. It could list, for example, a variety of sub-crisis
provocations which could involve a region in conflict. It would then seek a
complete list of each possible action and counteractions, and consider the
various paths of escalation and de-escalation. The most likely of these paths
are then selected for more careful consideration.
Approved For Release 1999/09/25: CIA-RDP85T00875R001100020002-2
 Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
N/A TIONAL (111AIRACTElt STUDIES-these draw upon the insights
of psychology, cultural anthropology, and sociology to study the effect of
national character, or the dominant personality traits of the adult members
of a society, upon individual and group political behavior within a state or
comparatively, between states. The study of national character was
stimulated by the growth of intere:it in non-Western societies where the m-
pact of social, cultural and psyc.lological variables on political systems
could not simply be assumed by the Western observer. (see: modal per-
sonality)
NETWORK ANALYSIS-u method for depicting the sequence and
interrelationship of actions, events, and benchmarks essential for progress
toward, and completion of, a significant goal or objective It provides a
means for tying together in reasonable order and sequence, the seemingly
infinite mix of variables that will influence an outcome-including
decisions, actions, and events. A network plan is a blueprint of the mental
logic of the analyst which consists of nodes representing events or
benchmarks, and arrows representing activities, sequence, and in-
terdependency. In political analysis, the analyst may employ network
analysis either retroactively to study the sequence which led to a past event
or decision once all the data is collected, or can hypothesize the probable
network leading to some future evert or decision and plug the data into it as
such data is received. This method can also be combined with others, such
as DECISION THEORYand CROSS-IMPACTANALYSIS.
GOVERNMEN
SHARE
INCiIEASED
CONSERVATIVE
WINS
FOREGN
INVESTGi1S
PULL OUT
FOREIGN AID
REPLACES
INVESTMENT
UNEMPLOY?
MENT
RISES
STABLE
ECONOMY
NETWORK ANALYSIS OF POSSIBLE OUTCOMES OF A HYPOTHETICAL ELECTION
OIL
COMPANIES
POLITICAL
APPOINTEES
REPLACE
EXPERTS
PRODUCTION
DECLINES
Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
 Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
NON-LINEAR PROGRAMMING-a mathematical problem-solving
technique which selects optimal solution subject to certain constraints. In
LINEAR PROGRAMMING, elationships between the variables must be
expressed by straight lines. Non-linear programming is designed to cope
with the more complex set of problems whose relationships are curvilinear.
Although there are general solution techniques for linear programming, not
all non-linear programming problems car,, be solved. (see: MATHEMATI-
CAL PROGRAMMING)
OPERA tfIONS RESEARCH-a generic term encompassing most of
the, Techniques defined in this glossary. Operations research is generally
take; to include almost any systematic application of scientific,
mathomaticai, or statistical methods to decision-making problems.
ORGANIZATION THEORY-as applied in political analysis, seeks
to explain political system3 and events in terms of the nature and functions
of the large, complex, goal-oriented groups that make up the political
system. Organization theory is concerned with the way organizations seek
goals and adapt to their environment, as well as with individual behavior oc-
curring within organizations. An organization theory approach to the study
of Soviet military decision-making, for example, would emphasize the
routine behavior, goals, and functions of the various significant groups
within the Soviet government and military. Some organization theorists
have postulated that all organizations behave in essentially the same way,
regardless of the society or culture in which they are found, and are
therefore essentially predictable. At present there is insufficient evidence to
support such a hypothesis. (see: BUREA UCRA TIC POLITICS)
PANEL STUDY-a techiiique for comparing the attitudes of an it
dividual over time, as contrasted with survey research techniques which
measure the attitudes of groups or aggregates of people at only one point in
time. Through interviews of the same people at two or more points in time,
the panel study permits the analyst to observe changes in individual at-
titudes that might be attributable to intervening events. The intervening
stimulus for example, might be a pub',,,- information program conducted
between the first and second interviews and planned as part of the study.
This kind of study is called an impact panel. (see: SURVEY RESEARCH)
Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
 Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
PARTICIPANT OBSERVATION-a research method by which the
investigator becomes a member, or poses as a member, of a group he is
studying. As a participant the researcher tries to become familiar with the
perspectives and assumptions of group members. As an observer, he tries to
view their behavior objectively and eliminate the group's values and biases
from his analysis. (see: GROUP THEORY)
PATH ANALYSIS-a statistical technique for dealing with a group of
interrelated dependent variables (those which the analysis is trying to ex-
plain or predict) for which there are several ultimate variables that corn-
pletely determine them. The ultimate variables may be observed, such as in-
come as a determinant of amount of tax paid, or they may be hypothetical
underlying variables-intelligence.as a determinant of IQ test scores.
For example, unemployment (1) and inflation rate (2) might be observed ul-
timate variables and popular dissatisfaction (a) and government repression
(b) the hypothetical ultimate variables related to the observed dependent
variables of riots (A) and political assassination (B). As shown in the
graphic above, the ultimate variables (1, 2, a, and b) completely determine
the observed variables (A a,'d B), as indicated by the one-headed arrows.
Variable B, however, is not only directly dependent on the ultimate
variables I and 2, but is also indirectly dependent on them through the in-
termediary of variable A.
Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
 Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
Once the relationships between variables have been identified, path analysis
specifies cc;uations which describe these rclutionshipv using coefficients that
measure the amount of variation in a particular dependent variable that is
directly attributable to anotl'cr variable. The coefficients in equations of
path analysis are called path eocr//Ic/e',tts. (see: CORRELATION, coef
1'ERCI;PTION ANALYSIS-a mode of decision-making analysis
which studies a decision-maker's or group's perceptions-that is, the
meaning assigned through the interaction of sensory impressions with
beliefs and values. The study of perceptions poses difficult problems of data
gathering because sensory impressions, beliefs, and values are internal to
each individual and must be inferred from observable behavior. For this
reason, perception analysis usually relies upon biographic sketches and
C'ONTE'NT ANALYSIS of speeches, statements and writings to reveal
perceptions. Among the factors that affect perception of political
phenomena are ideology, personality, past political activity and experience,
characteristics of' a particular decision stiuation, potential costs and
rewards, available information, and the emotional state of the perceiver.
(see: CONTENT ANALYSIS)
POLICY SCIEI\ :'E-an emerging field of inquiry concerned with the
application of scientific methods and insights to the improvement c,i public
policy-making and the solution of public problems. COST BENEFIT
ANALYSIS, ENDS-MEANS ANALYSIS, and DECISION THEOR Y
arc among the policy science techniques. Policy analysts are particularly
concerned with the process by which public policy decisions are made and
with clarifying goals, evaluating alternatives, identifying trends, and
analyzing the policy situation.
PSYCHO-BIOGRAPHY-a biographical technique which focuses on
the personality traits, development, attitudes, and beliefs of tl.e individual
under study in order to understand and predict his political behavior. (see:
SOCIAL/PSYCHOLOGICAL APPROACH)
Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
 Approved For Release 1999/09/25: CIA-RDP85T00875R001100020002-2
11SYCIIOLINGUISTICS-the study of language and communication
as it relates to the individuals who use the language. Psycholinguists argue
that language (everything we know) and speech (behavior expressing what
we know) play an intimate, vital, and complex role in human behavior.
They use the tools of both psychology and linguistics to study the mental
processes underlying the acquisition and use of language. Psycholinguistics
can be applied to either the study of an individual lender-what does the use
of ccrtu~n colloquialisms by a particular leader reveal about his personality,
for example-or to study the thought processes and behavior of ethnic or
national groups--how does the absence of verb tenses in the Chinese
language affect Chinese thought processes. Psycholinguists have shown
how language governs understanding and comprehension. Several African
languages, for example, use the same word for blue and black, and those
who speak only these languages. cannot dif'f'erentiate between the two colors
until shown the dif('erence.
QUANTIFICATION-the assignment of numerical values to political
variables so that they can be measured or compared by quantity.
Quantification of any political variable may be achieved, for example, by
the judguiental assignment of numbers to the concept being examined or by
find;ng numerical indicators to represent a political variable or concept that
cannot be easily quantified. Gross national product, which can be expressed
in terms of dollars, for example, can be used as one indicator of national
power. (see: SCALING)
QUEUE THEORY (waiting line theory)-a systematic study of the
various characteristics of the flow of units through some limited service
facility, involving congestion and delay. The input to a queueing study
would include data on the service system, rates of service, rates of arrival,
priorities, feedbacks, etc. Its output would include measurements of the
average lrngth cf queues, utilization of service facilities, comparisons of
diffcrunt priority ranking systems, effects of multiple servers, bottleneck
areas, etc.
REGRESSION-a common statistical technique for predicting a
dependent variable-the condition the analyst wishes to explain or
predict-from knowledge of its association with independent
variables-those variables used to explain changes in the dependent
Approved For Release 1999/09/25: CIA-RDP85T00875R001100020002-2
 Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
variable. Simple regression uses one variable to predict one other variable.
Multiple regression predicts the dependent variable from a combination of
other variables. The regression technique is designed to produce a mathe-
matical formula which describes the relationship between the dependent
variable and the independcr,t variable. For example, the number of rio!.s in
a city (dependent variable) may be related to the level of unemployment
(independent variable). The mechanics of simple regression produce a
straight line through the data points which has the "best fit." The criterion
for determining "best fit" is known as the least squares criterion which
minimizes the sum of the squared vertical distances from each point to the
regression life. A correlation coefficient is a measure of how close the points
are to the line. A perfect correlation of 1.00 means that each observation lies
on the line. The equation which describes the line is called the regression
equation and usually takes the genet al form: Y=a+b(X) where "Y" represents
the dependent variable, "X" represents the independent variable, "a" is the
point where the line intercepts the vertical axis, and "b" is a constant which
measures the impact of X upon Y. The value of b equals the slope of the
regression line. (s:c: CORRELATIONc?ast squares, slope)
REGRESSION
LINE
LEVEL OF UNEMPLOYMENT
RICHARDSON MODELING-a predictive method introduced by
Lewis F. Richardson's Arms and Insecurity which attempts to expla: ; arms
races between two nations or groups of nations. Richardson aFsumed alert
interaction between the two, modified by ambition, hostility, fatigue, and
cost. He evaluated stability criteria and equilibrium conditions, attempting
to predict arms expenditures by each sii. . This type Af interactive model
can also be applied to other competitive situations Such as retail sales,
military intervention, etc.
Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
 Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
RISK ANALYSIS-a simulation technique applied to problems of
decision making under uncertainty in which the probability distribution of
various possible outcomes is estimated, Risk analysis is most ef'f'ective in
comparing alternative policies-it demonstrates which one oi'l'ers the best
chance of high return, and how likely it is. Risk analysis generates the
expected values of various courses once the probabilities and outcomes have
been supplied.
SCALING-a method of measuring political variables that cannot be
counted in natural units, I f we are interested in analyzing domestic political
instability over a five-y;;ar period in Latin America, Ior example, the events
which indicate instability could be valued from 0 (stable) to 6 (extreme
instability) as follows: O-General election, I-Resignation of a cabinet
official, 2-Peaceful demonstration, 3-Assassination of a significant
political figure, 4-Mass arrests or violent riots, s-Coup d'etat, 6-Civil
war. The number of times each event occurred per year would then be
multiplied by the value of the event to produce scores which, when added
together, would equal the yearly score. The yearly scores could then be
plotted on a graph or chart of domestic instability. (see: GUTTMAN
SCALE)
SCENARIO GENERATION-a narrau ?e description of a sequence of
events which illustrates how a future situation might logically evolve.
Scenario writing is a representation of possible futures given certain basic
initial assumptions. The major emphasis is usually on the subtle points
where assumptions may significantly change the projected outcome. The
use of scenarios is especially applicable to planning or normative
forecasting, as it provides a mechanism for considering the consequences of
alternative policies and for sharpening the analyst's thinking about options,
generalizations, and probabilities.
SEMANTIC DIFFERENTIAL-a technique for measuring the
connotative meaning that people assign to concepts. It consists of several
scales ranged betwr;en two polar adjectives (good-bad, strong-weak,
active-passive, for example) and one or more concepts to be rated on the
scales. The semantic differential tool is used in political science for research
on political attitudes and beliefs. Voters for example might be asked to rate
their attitudes toward political issues or candidates on such scales in order
to determine the difference in what those issues and candidates connote to
people with different political beliefs. (see: attitude scaling)
Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
 Approved For Release 1999/09/25: CIA-RDP85T00875R001100020002-2
SI,NSI'1'IVI'I'Y ANALYSIS-the process of testing it model or
simulation in order to determine whether relatively minor ch ng,as in the
original assumptions cause significant variation in the results, II'the output
of the system is subject to wide swings in response to small changes of a
single variable in the input (i.e., the model is highly sensitive) then the
starting data must be specified with great care. II', however, the results are
less volatile in following the fluctuations of the input (i.e., low sensitivity)
then more error may be permitted and we can still retain confidence in the
model's validity.
Sk,QUENTIAL ANALYSIS-the analysis of' data derived by a
sequential method of' sampling in which the members of the sample are
drawn one by one or in groups in order, and the results of the drawing at any
stage decide whether sampling is to continue. The sample size is not fixed in
advance, but depends on the actual results and varies from one sample to
another. The sampling terminates according to predetermined rules
according to the degree of precision desired. (see: sample, population,
SURVEY RESEA RCH)
SIMPLEX METHOD-a computational linear programming
technique designed for solving problems which contain many variables or
constraints. In such complex problems, graphic techniques are inadequate
and manual computation is too time consuming. The simplex method
avoids searching through all possible alternatives in a trial-and-error
approach, by using an acceptable solution as a criterion for Testing the
optimal solution.
SIMULATION-the reproduction, in it laboratory or experimental
situation, of selected aspects of a real or hypothetical system? Political simu-
lations are usually simplified or scaled-down representations of political
behavior, such as decision-making. Some simulations involve only human
iateraction; others may be done by computer; and some combine both
features. Computer simulation requires complete specification of the
relevant variables and decision rules. With human participants, at least
some of the rules governing the interplay of model components are left un-
specified. If only human participants are involved, the technique is usually
called gaming.
Approved For Release 1999/09/25: CIA-RDP85T00875R001100020002-2
 Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
SOCIAL/PSYCIIOLUCICAL APPROACH-a mode of analysis
which focuses upon human personality and behavior as a political
determinant, The general approach encompasses a variety of techniques
and methods including PSYC'/lO.IJIOGRAPIIY, CONFLICT
BEHAVIOR ANALYSIS, NA77ON11L C/IARACTER and modal
personality studies, psychoanalysis, DECISION-MA KING,
CROSS-CULTURAL, and ELITE ANALYS'IS.
SOCIOMF:I'RY-a technique for the analysis of relationships within
small groups. Sociometry uses data from questionnaires, interviews, or
expert opinion to determine which persons most associate in a given activity
with other persons of the group. Patterns of interaction between group
members-such as power relationships and personal and ideological
preferences-are revealed. Diagramming the power relationships between
the members of a cabinet that is divided between radicals and moderates,
for example, reveals the pivotal position of the Finance Minister, Mere
rank order listing of the government's leaders by degree of political power
often conceals the importance of the individual as well as much information
about relationships between individuals. The graphic representation of this
information by meads of circles, connecting lines, or other geometric
figures, is called a sociograin. (see: attitude scaling, sociogram)
SoclometrIcally-mapped diagram
Moderates Radicals
2 7 Noy
4/ N5" \9
1 President
M.mMler? of
2. ror.Ign Allal,,
3. Security
4. Ddense
6 Flnence
O, Education
7. Dereloamenl
6. Slate
D. Culture
S'T'IMULUS-RESPONSE ANALYSIS-(see: INTERACTION
ANAL YSIS)
STRUCTURAL-FUNCTIONAL ANALYSIS-political analysis that
focuses on the functions performed within a political system and the
structures through which they are performed. A fiunction, this context, is
the result or consequence of actions which satisfy the needs or demands of
the system for its perpetuation. A structure is ai"y pattern which governs
all.;:ions, including some concrete organizational structures as politic- I
parties, legislatures, or juntas. Structural-funct,,onaf analysis post:.. _s
that certain necessary functions must be performed if a system is to persist.
It provides a framework for analyzing systems, stressing the
interrelationships between structures and functions within a given system.
One of the most influential applications of structural-functional analysis to
Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
 Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
political science wits by Gabriel Almond who provided it model of five
political or input functions-political socialization, political recruitment,
interest articulation, interest aggregation, and political communication--
and three governmental or output functions -rule making, rule applica-
tion, and rule adjudication. (see: FUNCTIONALISM)
SUBSTITUTION ANALYSIS-a technique for technological
forecasting which measures the percentage or a given market that has been
or will he .;ssumcd by a new technological innovation. Typically, the data is
represented as it substitution curve over time, and when the new process has
reached it level of 15 percent of the market, the trend is considered
irreversible. Extrapolation to the future is then possible. For example, in
forecasting a country's shill to curtain high performance weapr, ns, we
measure the trends in tnc percentage of the total "market" already c t hared
by the new HPW's and project forward.
SURVEY RESEARCI--the investigation of social phenomena by
means of interviews and questionnaires administered to a number of
respondents for the purpose of studying people's attitudes, opinions,
behavior, and social attributes (such as age, sex, political affiliation, and
economic class) and the relationships among these characteristics. Survey
research ordinarily entails the study of samples, from which inferences may
be drawn about the whole population under investigation, and provides data
on the characteristics of each individual interviewed. This contrasts with
aggregate data, such as trade or census statistics, which summarize the
characteristics or transactions of a whole population. (see: sample,
population, aggregate data)
SYSTEMS ANALYSIS-in political analysis, a method of analyzing
politics as a set of interactions occurring within, but analytically distinct
from, the larger social environment. A, system, in this context, is a patterr
of interactions or persisting relationships among political enti'aies. A system
that constitutes an element or a larger system is called a subsystem-for
example the legislature within a government, or he committee structure
within a legislature. The setting within which a system occurs is its
environment. Political analysts have directed systems analysis to such
questions as the relationships of political systems to their social and
economic environment, the persistence of systems over time, system
stability, system functions and related structures (the concern primarily -?f
STRUCTURAL-FUNCTIONAL ANALYSIS), the identificatic,i,
classification, and measurement of system inputs and outputs, and
communication within a political system (COMMUNICATIONS
T!IEOR Y).
Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
 Approved For Release 1999/09/25: CIA-RDP85T00875R001100020002-2
In an organizational-management context, systems analysis refers to the use
of a systems framework to evaluate strategies, alternatives, and program
objectives with the object of using resources most effectively and achieving
goals. It provides decision-makers with a full, accurate and meaningful
summary of the information relevant to any decision, by charting,
describing, and reducing to mathematical language any enterprise,
organization, or issue. The newly constructed "system" is then operated by
feeding in hypothetical data; by changing policies or circumstances and
revising the equations to fit the new assumptions. Further simulation
produces another set of'results, and successive "runs" of the process are then
compared to attain the optimum configuration of'elements.
SYSTEMS DYNAMICS-a general name for a variety of modeling
approaches to long range forecasting based upon estimation of the levels
and rates of'change of key variables. Systems dynamics does not necessitate
as sharp a degree of precision in the original assumptions as do other
modeling methods.
TECHNOLOGY ASSESSMENT-the systematic study of a new
technology on a society with special emphasis on impacts that are
unintended, indirect, and delayed. Essentially the steps are to define the
scope of inquiry, outline the state of the art in the technology being assessed
(as well as in related technologies), develop state-of-society assumptions
regarding non-techno;ogical factors, identify impact areas (law, medicine,
and social relationships, for example), and identify and analyze the options
open to the society.
TIME SERIES ANi LYSIS-a general name for a number of
processes using past experience alone to predict. future values of a variable
which fluctuates over time. The result is a predictive formula which best
fits the observed data. It can represent trends, cycles, seasonal variations,
etc. The most sophisticated time series analyses are similar to a
REGR?ESSION equation in which the dependent variable is the prediction,
and the independent variables are the past values the variable itself has
previously taken. The predictive ability of the equation can be tested on
either past or new data. For example, enemy infiltration into a country may
Approved For Release 1999/09/25: CIA-RDP85T00875R001100020002-2
 Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2
have a general upward trend, modified by seasonal cycles due to weather.
Additionally, the rates will be affected by policy choices on both sides, and
there may be it simple random component. Time series analysis would
attempt to sort out all the different elements and provide a predictive
fore.iIa.
Diachronic analysis and longitudinal analysis are other names applied to
studies having a time dimension. Research dealing with data drawn frnm a
single time period, or examined without reference to time as a variable, is
called synchronic or cross-sectional analysis.
TRANSACTION ANALYSIS-(sec: EVENTS ANALYSIS)
VOTING BEHAVIOR STUDIES-(see: ELECTION ANALYSIS)
WAR GAMING-a simulation of a military battle consisting of two
opposing forces with a referee or control team. The control team evaluates
the impact and outcome of the decisions of each side and writes the scenario
for successive plays. The game may be actually played out on a field, or
programmed and run by computer. By repetition and experimentation, this
type of simulation may be used to "optimize" military resource allocations,
select strategies, and test innovations.
Approved For Release 1999/09/25: CIA-RDP85T00875RO01100020002-2