A SURVEY OF THE COMMUNITY'S USE OF NEW ANALYTICAL METHODS

• Document Type: CREST
• Collection: General CIA Records
• Document Number (FOIA) /ESDN (CREST): CIA-RDP83M00171R001800110002-5
• Release Decision: RIPPUB
• Original Classification: S
• Document Page Count: 52
• Document Creation Date: December 15, 2016
• Document Release Date: August 1, 2003
• Sequence Number: 2
• Content Type: REPORT

Approved For Release 2003/09/02 : CIA-RDP83M00171 R001800110002-5
A Survey of the Community's Use
of New Analytical Methods
DOS, OSD and DIA review(s) completed.
Approved For Release 2003/09/02 : CIA-RDP83M00171 R001800110002-5
 Approved For Release 2003/09/02 : CIA-RDP83M00171 R001800110002-5
Page
Synopsis
Introduction
The Promise of New Techniques
The Central Intelligence Agency
Office of Current Intelligence
Office of Political Research
Office of Strategic Research
Office of Economic Research
Office of Geographic and Cartographic
Research
Office of Research and Development
The Analytical Support Center
Office of Training
The Defense Intelligence Agency
29
Directorate of Intelligence
29
Directorate of Estimates
30
Directorate of Collection
33
Defense Intelligence School
34
The Defense Advanced Research Projects Agency
35
The Department of State
39
Approved For Release 2003/09/02 : CIA-RDP83M00171 R001800110002-5
 Approved For Release 2003/09/02 : CIA-RDP83M00171 R001800110002-5
Synopsis
The Intelligence Community is actively seeking to
develop and apply promising new analytical techniques.
Community agencies already use a number of new analytical
methods and are considering and experimenting with others.
Interest in these methods varies from agency to agency and
from office to office, the differences due in part to the
greater susceptibility of some kinds of work to quantitative
treatment than others.
Both the Central Intelligence Agency and the Defense
Intelligence Agency routinely use sophisticated analytical
methods in their studies of military forces and weapons.
CIA has, in addition, applied new techniques to political,
economic, and cartographic research. In both agencies,
formal courses are held to introduce analysts to new methods.
The Department of State, whose work lends itself less easily
to the use of quantitative techniques, has commissioned
comprehensive studies of the usefulness of new analytical
methods (and of other developments in the social sciences)
to political research. The Advanced Research Projects Agency
of the Department of Defense is applying new techniques to
several fields of interest to the Intelligence Community,
including forecasting and early warning.
1
Approved For Release 2003/09/02 : CIA-RDP83M00171 R001800110002-5
 Approved For Release 2003/09/02 : CIA-RDP83M00171 R001800110002-5
It would be premature to judge the value of many of
the methods which are described in the body of this report.
Clearly, however, the exploration of new methods of analysis
is a valuable Community exercise. More important than any
particular new technique is this general willingness to
think critically about the way in which analysis is conducted,
to consider whether there are not more informative ways of
ordering the same information, whether a tried analytical
method is in fact applicable to novel problems, and whether
traditional methods are best suited to the interpretation of
information from new kinds of sources.
Approved For Release 2003/09742 : CIA-RDP83M00171 R001800110002-5
 Approved For Release 206& /~ EdIA-RDP83M00171 R001800110002-5
A Survey of the Community's Use of New Analytical Methods
Introduction
"New analytical methods" is a notoriously imprecise
term, one that has at times been applied without much
discrimination to techniques that are neither new nor,
strictly, analytical. This paper compromises with that
imprecision. On the one hand, the methods that it surveys
are analytical ones; they are, that is, explicit and
systematic ways of ordering and interpreting facts, and
are not simply devices or arrangements for storing, trans-
mitting or presenting information, however necessary and
helpful these are. Construed in this way, furthermore,
analytical methods are not necessarily dependent on a com-
puter, and the use of a computer is here considered neither
necessary nor sufficient evidence that an analytical method
is being applied.
On the other hand, one must be somewhat less demanding
with respect to the novelty of these methods. While their
use in intelligence (particularly in political analysis) is,
with some important exceptions, recent, many of these methods
have been used for some time by political scientists,
This paper is based on discussion with analysts
throughout the Community who are now involved in the
development of new methods, on written precis of their
work which some helpfully provided and on a summary of
its efforts in this field by CIA's Directorate of Intelli-
gence. ("Progress on New Methodologies in the DDI," 21
August 1974.)
Approved For Release 20SftREMA-RDP83M00171 R001800110002-5
 Approved For Release 20LZ*~E ETIA-RDP83M00171 R001800110002-5
economists and corporate managers. It is, of course, only
their application to intelligence that is of concern here,
and in that application many of them are indeed new.
As the phrase is used here, "new analytical methods"
refers only to those techniques designed for political,
economic, and military intelligence. But for the same
reason that a discussion of new techniques for scientific
analysis is excluded from this study--because such methods
are less apt to represent a fundamental departure from
previous ones--it is something of a distortion to speak of
new analytical methods in economic intelligence, a field in
which quantitative methods have long been used and in which
the need continually to adapt methods to new emphases and
new subjects seems well accepted. This is reflected in the
rather brief treatment given here of the work of the Office
of Economic Research.
The Promise of New Techniques
New methods should enhance analysis in several ways.
When they are appropriately used, they demand, if not the
elimination, at least an explicit expression of subjective
judgments, which are then more readily weighed and amended.
Where information can, without distortion, be expressed
numerically, these methods permit greater precision and the
systematic consideration of a great many more influences
on an event than would be possible without them. Indeed,
SECRET
Approved For Release 2003/09102 : CIA-RDP83M00171 R001800110002-5
 Approved For Release 2005 RECTA-RDP83M00171 ROO1800110002-5
some problems are so complex--like the hypothetical ex-
change of nuclear weapons, the study of the influences on
the world's trade in a certain commodity, or the causes of
political violence--that some explicit and, if it is possible,
quantitative method of analysis may be necessary for genu-
inely comprehensive treatment.
New problems may require the revision and even the
abandonment of previously useful approaches and the creation
of original ones; the development of "military economics"
is an example. Similarly, new sources of intelligence may
demand new methods of interpretation and analysis, particu-
larly if the sources are prolific and are to be exploited
fully; instances are the diverse ways in which satellite
photography is being exploited. But perhaps the most
important if not the most obvious fruit of the development
of these new methods is the opportunity it provides to step
away from the actual work of analysis and to examine the
assumptions which have guided it.
Despite these merits, an unequivocal recommendation
of new methods is not intended. The belief that all or
even most of the many new techniques developed for, say,
political analysis can be successfully adapted to intelli-
gence is unfounded. Many such methods demand a great deal
more information than is normally accorded analysts. More-
over, while not all new techniques are quantitative ones,
SECRET
Approved For Release 2003/09/02 : CIA-RDP83M00171 R001800110002-5
 Approved For Release 20~i9J102`.'TIA-RDP83M00171 R001800110002-5
many are, and some issues simply do not lend themselves
to numerical treatment.
Even when these reservations do not apply, even
when a new analytical method might in principle be applied
to intelligence, it may be evident after some examination
that it is unnecessary or unwise to do so. Unnecessary
because in many instances conventional forms of analysis
will be quite sufficient. Unwise because the particular
method may require more time or more specialized analytical
skills than an office can be expected to provide, and more
sophistication and more patience than readers can be ex-
pected to bring to bear.
Much of the Community's use of new analytical methods,
particularly for political analysis, is still tentative and
experimental, and appropriately so. These investigations,
furthermore, are often done at various removes from the
actual production of intelligence, and that is probably
appropriate as well. Certainly some freedom from the
insistent press of business is needed for the considered
adaptation of promising methods to analytical problems,
even in offices in which the use of quantitative methods
is already common. The establishment of groups within
analytical offices to concentrate on such experimentation
and the letting of contracts for the same purpose offer
at least the possibility of this freedom.
SECRET
Approved For Release 2003/09402 : CIA-RDP83M00171 R001800110002-5
 Approved For Release 20MRE IA-RDP83M00171 R001800110002-5
Nevertheless, the importance of actively involving
analysts in these investigations and early applications
should always be borne in mind. Analysts should be able
to ensure that any methods developed or adapted for intelli-
gence promise real benefits. New methods should also be
introduced in such a way that the sophistication of analysts
with regard to these techniques is increased, a benefit
not likely to be realized if new methods are developed or
used entirely apart from the analysts, whether by contractors
or by small groups within the Community with only formal
links to other analysts. Some balance needs to be struck,
then, between imposing additional burdens on already busy
analysts and subjecting this preliminary work to the risk
of sterility.
What follows is a survey of the development, the
adaptation and in some instances the already routine use
of new analytical methods by the Defense Intelligence
Agency, the Central Intelligence Agency and the Bureau of
Intelligence and Research of the Department of State.
Several projects which the Defense Advanced Research Projects
Agency is conducting are sketched as well.*
Although the National Security Agency has devised
sophisticated methods for using computers in the analysis
of signals intelligence, it is not now experimenting with
the kinds of analytical methods that are treated here.
SECRET
Approved For Release 2003/0 102 : CIA-RDP83M00171 R001800110002-5
 Approved For Release 20089?4 i -RDP83M00171 R001800110002-5
THE CENTRAL INTELLIGENCE AGENCY
Office of Current Intelligence
The Office of Current Intelligence has charged its
Methods and Technology Staff with acquainting analysts in
OCI with new analytical methods. The Staff encourages
the use of those qualitative and quantitative techniques
that its own studies suggest are applicable to intelligence.
As part of its educational work, the Staff publishes
a newsletter, Notes on Methodology, each issue of which
describes the basic principles and possible uses of a
promising technique. Some of the articles have treated
specific analytical problems, for whose solution novel
methods have proven useful. Notes on Methodology is
widely distributed within CIA and, because it is an in-
formal publication, can be prepared economically. In
conjunction with the newsletter, the Staff has carried
out a series of briefings to educate analysts in the use
of new techniques. It has also published a comprehensive
Glossary of Terms and Techniques for Political Analysis
a unique and particularly helpful guide through the some-
times obscure language of these methods.
OCI has furthered the education of its analysts in
these techniques in a variety of other ways, enrolling
SECRET
Approved For Release 2003/09/02 : CIA-RDP83M00171 R001800110002-5
 Approved For Release 2~!9t lA-RDP83M00171 R001800110002-5
officers in courses offered by the Office of Training
and sending analysts to attend academic conventions on
this subject. In addition, the office recently sponsored,
in conjunction with the Office of Research and Development,
a symposium on the subject of
The par- 25X16
ticipants in the symposium included analysts not only from
OCI but from other parts of CIA and from the State Depart-
ment and several universities as well.
The Methods and Technology Staff also assists in
applying new analytical methods to current problems, and
of course such efforts also serve to acquaint the current
intelligence analysts involved in the project with whatever
technique is being employed. The Staff's primary intention.
in this respect is to tailor new methods and new ways of
presenting analysis to the particular demands of current
intelligence. The Staff emphasizes approaches which do not
require lengthy preparation, specialized data, or the ex-
25X16 tensive use of computers. The fruits of this emphasis
SECRET
Approved For Release 2003/09102 : CIA-RDP83M00171 R001800110002-5
 251Bved For Release 2 -mfoR TIA-RDP83M00171 R001800110002-5
OCI has begun work on a number of political models
which should assist current analysts in refining their per-
ceptions of a country's politics and of relations between
countries. A preliminary version of a model representing
conflict over sources of energy, both within a country and
internationally, has been completed for OCI byl
25X1A
(ORD providing most of the funds). The model is designed
to point out specific deficiencies in our knowledge of such
conflicts and to refine our understanding of the complexities
of international competition.
, is also assisting 25X1A
OCI in the preparation of a quantitative analysis of ter-
rorism in Argentina.) Modeling and simulation are techniques
which may be of particular use when a number of OCI analysts
are dealing with various parts of a larger issue.
The Methods and Technology Staff recently made a "UN
Voting File" available to analysts in OCI. Designed by
the Political Science Department of the US Naval Academy
and the Office of Management Systems of the.Department of
SECRET
Approved For Release 2003/08902 : CIA-RDP83M00171 R001800110002-5
 Approved For Release 2003SCO RDP83M00171R001800110002-5
URET
State, the file allows analysts to obtain quickly the voting
records both of individual members and of certain groups
within the United Nations. It also permits detailed com-
parisons of such records.
It is important to note, however, that while OCI is
experimenting with a number of promising new methods, the
office has reservations about the usefulness of some tech-
niques for current intelligence. The office believes that
much of the information with which it works cannot be ex-
pressed numerically without grave distortion and that, in
any case, some of these methods demand more time and prepa-
ration than the character of the office's work allows.
Office of Political Research
The Analytical Techniques Group (ATG) within the Office
of Political Research is responsible for the application of
novel, frequently statistical approaches to the analysis
of political intelligence. Its work includes the adaptation
of existing techniques used in universities and private
industry, as well as the creation of new techniques and ar-
rangements to suit the unique needs of intelligence production.
In the nearly two years since its creation, ATG has achieved
some of its most fruitful results in developing ways of
bringing the opinions of a number of experts to bear upon
selected subjects in a rigorous manner, and in using new
devices for the graphic communication of finished intelli-
gence.
SECRET
Approved For Release 2003/0902 : CIA-RDP83M00171 R001800110002-5
 Approved For Release 20 JRE IA-RDP83M00171 R001800110002-5
The following list of projects illustrates the scope
of OPR's work in these new methods:
a. Beginning in December 1973 the Analytical
Techniques Group applied the Bayesian technique of
analysis in asking a group of experts to assess "The
Likelihood of a Major North Vietnamese Offensive."
This exercise at first involved only Agency analysts.
After the favorable reception of the study's
periodically published reports, it was expanded in
February 1974 at the request of the DCI and the
United States Intelligence Board to include experts
from several agencies in the Intelligence Community.
Both phases of this Bayesian analysis were concluded
in June 1974, and an evaluation was published shortly
thereafter.
b. Since June 1974 the Office has conducted
Bayesian analyses of two topics: "The Likelihood of
Sino-Soviet Hostilities" and "The Likelihood of Arab-
Israeli Hostilities." In each case the analysis has
been designed to examine a range of possible outcomes
beyond the basic problem of whether there will be a
war. Reports are issued every six weeks for the Sino-
Soviet study and biweekly for the Middle East project.
c. Based on its experience in these projects,
the Analytical Techniques Group has published a hand-
book on the application of Bayesian analysis to intel-
Approved For Release 2003/aTOc TRDP83M00171 R001800110002-5
10
 Approved For Release 20Q beREYIA-RDP83M00171 R001800110002-5
ligence. This publication was put together to explain
the technique and to encourage its use as a device
which would complement traditional ways of looking
at political or strategic phenomena. The group has
also developed computer routines in various machine
languages which will assist an individual or group in
the conduct of a Bayesian exercise.
d. ATG has developed an operational analytical
model for assessing the causes and consequences of
political violence. The first phase of this project
involved the establishment and adaptation of the
model, developed The model 25X1A
was tested in a historical case study of the Chilean
coup to determine its usefulness in gauging situations
of potential political conflict. The second phase,
now in progress, is designed to test further the
diagnostic and predictive powers of the model.
Specifically, it involves the application of the
model to certain countries--Ethiopia, Thailand and
Argentina--in which political dissension and popular
dissatisfaction appear to be significant, and the
publication of monthly analytical reports based on this
use of the model.
e. The office has also staged a number of political
games as a structured way of educating analysts and of
SECRET
Approved For Release 2003/09/02 : CIA-RDP83MOO171 ROO1800110002-5
11
 Approved For Release 2009 i -RDP83M00171 R001800110002-5
inducing them to consider some of the assumptions
which are implicit in traditional analysis. Games,
which are formal devices for exposing and testing
the decisions of participants who play the roles of,
for example, the leaders of competing states, have
been run on the following subjects: (1) the dispute
involving the Iraqis, the Iranians, and the Kurds;
(2) the potential for conflict on the border separating
Chile and Peru; and (3) the conflict between the
decision-making processes of the competing political
groups in Brazil. Consideration is being given to
running games on Southeast Asia and on Chinese in-
ternal politics.
f. An exercise employing Delphi techniques to
study the future of the Portuguese African territories
was held in the spring of 1974. Delphi--a very popular
analytical technique in private industry--employs a
process of repeated responses and reevaluations to
identify areas of clear agreement or disagreement on
a problem among experts. While it has not often been
used alone as an approach to' intelligence analysis,
its organizational features are quite applicable to
a system in which consultation and coordination among
analysts are common; accordingly, some aspects of this
process have been incorporated into the exercises which
ATG has developed.
SFCRET
Approved For Release 2003/09 : CIA-RDP83MOO171 ROO1800110002-5
 Approved For Release 2003/ aa[[ ~9f9~ DP83M00171R00125X18001100B 02-5
h. Other projects now under way or planned to
be undertaken soon will employ decision theory, cross-
cultural perception analysis, and methods for designing
scenarios for use in policy planning and analysis.
Office of Strategic Research
The Office of Strategic Research has for some time
used quantitative techniques in describing and estimating
the size, quality, deployment and costs of foreign military
forces and weapons. Its Planning Staff is specifically
responsible for encouraging and supervising the trial and
adoption of new methods. The Office has, however,
SECRET
Approved For Release 2003/09/1 : CIA-RDP83M00171 R001800110002-5
 Approved For Release 2c 8 f lA-RDP83M00171 R001800110002-5
25X1 B
successfully relied upon individual analysts to initiate
experiments and has found that its analysts, faced, for
example, with the need to predict the results of hypothetical
but enormously complex situations (e.g., an exchange of
strategic weapons), continually propose new applications
of quantitative methods.
25X1 B
bureaucratic processes and techniques drawn from organization
theory to its studies of the Soviet bureaucracy's perceptions
of US military forces.
The Center has drawn up a detailed program for the
refinement of existing methods and for the development of
new ones, a program in which OSR will be assisted by ORD.
The latter has, for example, agreed to support the design
particular location. Also planned is the improvement of
techniques which are already in use, such as a model of
SECRET
Approved For Release 2003/09i12 : CIA-RDP83M00171 R001800110002-5
 Approved For Release 20&FCR.IIA-RDP83M00171 R001800110002-5
25X1 B
have not yet been launched. The Center will undertake as
well the education of its analysts in the employment of
both tried and novel techniques.
Other divisions in OSR have also bent new methods to
their work. The Theater Forces Division, for example, 2
5X18
25X1 B
Working with the Office o
f
25X1
Joint Computer Support, the division has,
25X1 B
prepared, as part of this exploration, a model for assessi
ng
With ORD,
25X1A
25X1
25X1
in devising new approaches to the analysis of
the various influences on the course of
In addition, appraisals by the Theater Forces Divisio
of the conventional forces of the Soviet Union, the Warsaw
Pact and NATO rely on statistical tools which often requir
n
e
25X1
the use of computers and particularly the use of two filin
g
SECRE1
Approved For Release 2003/09/02 : CIA- DP83M00171 R001800110002-5
15
 25X1 B Approved For Release 2003/09/02 : CIA-RDP83M00171 R001800110002-5
Approved For Release 2003/09/02 : CIA-RDP83M00171 R001800110002-5
 Approved For Release 2003/09/02 IA-RDP83M00171 R001802tV9
Despite the successful and in many instances the routine
use of these analytical methods, the Office retains a certain
skepticism about their more general application to the kinds
of studies for which it is responsible. Like analysts in
OCI and other parts of the Community, analysts in OSR
believe that quantitative approaches used effectively for
Approved For Release 200 EIA-RDP83M00l71 R001800110002-5
 Approved For Release 2003/02C?f P83M00171 R001800110002-5
OL1p
academic studies often demand more information, time and
preparation than are available to an intelligence agency.
The paucity of information is especially a problem as,
unlike an academic researcher who can in some ways choose
a topic to suit a particular method, the analyst's "problem"
and the information he or she can draw upon are both largely
fixed. Accordingly, a technique must be adapted to a subject,
if that is possible, and often it is not.
Office of Economic Research
The Office of Economic Research has, as noted previously,
long used a variety of quantitative techniques in its analysis.
Such methods are of course more readily applied to economic
data than to information about, say, politics, and indeed
economists have made sophisticated statistical descriptions
and mathematical models an expected feature of their work.
In a sense, then, the phrase "new analytical methods" has
rather a different meaning for this office than it does for
some others, for in OER it is apt to suggest refinements
of quantitative approaches that already enjoy a general
acceptance in economic research.
Still, such methods are often designed for the study
of problems on which a great deal more information is
available than is often true of issues in economic intelli-
gence. The Office must accordingly evaluate these methods
for their usefulness to intelligence and then frequently
modify them. In 1968, the Office established the Systems
SECRET
Approved For Release 2003/091Q : CIA-RDP83M00l71 R001800110002-5
 Approved For Release 2003SQ R 1a-RDP83M00171 R001800110002-5
25X1 B
25X1 B
Development Staff (recently made part of the Development
and Analysis Center) to oversee this modification and
adaptation of established methods to economic intelligence
and to experiment with new ones as well. The working
relationships enjoyed by members of the staff and other
analysts in OER with economists in universities and other
parts of the government aid in this task of devising and
applying advanced techniques.
OER uses established econometric models in its work
as a matter of routine. In the last several years, the
Efforts by OER to develop and apply its own new ap-
proaches have also been fruitful. One project employs
To encourage and facilitate the use of new methods,
the Systems Development Staff has provided formal courses
and individual instruction to analysts from OER and from
SECRET
Approved For Release 2003/09I(2 : CIA-RDP83M00171 R001800110002-5
 Approved For Release 2MTVIA-RDP83M00171 R001800110002-5
25X1 B
25X1 B
other parts of the CIA as well. The Staff has trained
analysts both in the employment of the techniques them-
selves--courses in Bayesian analysis, macroeconomic model-
ing and trade simulations have been provided--and in the
application of computers to analytical problems.
Office of Geographic and Cartographic Research
The Office of Geographic and Cartographic Research
is pursuing a number of programs in which new methods are
being tested and then applied to the office's special
analytical concerns. In developing and experimenting with
new methods OGCR has sought the cooperation of OJCS, OSR,
ORD, IAS and NPIC and that of experts in other government
agencies, universities, and commercial firms. Supervising
the application of new techniques is the responsibility
of a special assistant to the Director.
With the assistance of academic and industrial
consultants, the Office has developed new methods for ex-
ploiting satellite photography. Two techniques have been
Similar methods may have
promise in the study of such diverse topics as water re-
sources, urban change, and the production of narcotics.
Analysts in OGCR have devised techniques for using, in
SECRET
Approved For Release 2003/091 : CIA-RDP83M00171 R001800110002-5
 Approved For Release 2003/OQV CRI rDP83M00l71R001800110002-5
25X1 B
25X1 B
geographic and strategic intelligence, information gained
The Cartography Division's accomplishments in develop-
ing automated cartographic systems should be noted for
they do permit the manipulation as well as the storage and
presentation of information.
is one such system
25X1
25X1
for producing maps that has become an important analytical
developed by OGCR. The Division is examining computer
mapping systems in use in universities and industry for
their possible application to intelligence.
The Geography Division is studying the possibilities
of a computer system for storing, retrieving and manipulat-
ing geographic information. Representatives from ORD,
OJCS, OSR, and OGCR are now reviewing the demands that could
be made of such a system.
The Special Research Branch of the Geography Division
has experimented with new statistical methods as part of a
program to provide intelligence to the National Security
Council for its use in preparing for the Law of the Sea
negotiations. Using studies prepared by the Center for
Naval Analysis, the Branch has employed
SECRE
25X1 B
Approved For Release 2003/09~j2 : CIA-RDP83M00171 R001800110002-5
 Approved For Release 2003N Ell-RDP83M00171 R001800110002-5
1B
vote on various issues. OGCR plans to extend the use of
these methods to such other subjects as world population
and food problems.
Office of Research and Development
In 1973, the Office of Research and Development es-
tablished the Center for the Development of Analytical
Methodology (CDAM). The Center is responsible for de-
vising new methods of analysis and adapting both these
original techniques and pertinent existing ones to the pur-
poses of intelligence. The members of the Center's staff
have backgrounds in various fields and are thus able to
work closely with analysts throughout CIA.
Although most of CDAM's work is in the areas of
science and technology, it has also contributed to the
development of new methods for economic and for political
research. The Center assisted, for example, in devising
SECRET
Approved For Release 2003/09tg : CIA-RDP83M00l71 R001800110002-5
 25X1 B
25X1 B
Approved For Release 20Q3iWO IA-RDP83M00171 R001800110002-5
the Center is investigating the value to analysts of a
model of international conflict over sources of energy.
This model emphasizes the political aspects of such con-
flict and, when it is refined, will assist analysts in
assessing the policies nations may pursue in response to
shortages of energy.
Analytical Support Center
The Center for the Development of Analytical Method-
ology also supervises the work of the Analytical Support
Center. Established in November 1974, and sponsored by
CIA, the Intelligence Community Staff and ARPA, the
Analytical Support Center was created to permit analysts
to experiment with new analytical methods which hold some
promise of directly assisting them in their work. The
contract for the Center was awarded
members of which bring academic qualifications in diverse
disciplines to the Center's work.
The Center will concentrate initially on the develop-
ment and application of analytical techniques for the study
25X1A
Approved For Release 2003/0 C -CIA-RDP83MOO171 R001800110002-5
 25X1 B
25X1 B
Approved For Release 2003T097P2~CIA-RDP83M00171 R001800110002-5
topics are selected not only for their importance to the
Intelligence Community but also for the extent to which new
techniques developed in the course of a study can be ap-
plied to other analytical problems. An important pre-
liminary to each study will be a survey of the academic
literature on the topic, to insure both that the study
will not merely duplicate a previous one and that there
is enough information on the subject to make an investiga-
tion fruitful.
The Analytical Support Center also holds symposia to
acquaint analysts with the merits of new analytical ap-
proaches. Techniques for analyzing
treated in one symposium, the study
Office of Training
The Information Science Center of the Office of Train-
ing runs the Information Science Training Program, which
is open to personnel from the entire Community. The
program, now in its eighth year, emphasizes the actual use
of new analytical approaches and computer systems by
SECRET
25X1 B
25X1 B
Approved For Release 2003/0982 : CIA-RDP83M00171 R001800110002-5
 25X1 B Approved For Release 2003/09/02 : CIA-RDP83M00171 R001800110002-5
Next 1 Page(s) In Document Exempt
Approved For Release 2003/09/02 : CIA-RDP83M00171 R001800110002-5
 Approved For Release 2003/(KC RL'6P83 R001800110002-5
The Center for the Study of Intelligence of OTR's
Intelligence Institute began in August a series of seminars
in which experienced analysts examine the conduct of
Approved For Release 2003/OSECBE-IDP83M00171 R001800110002-5
25X1
 Approved For Release SU9R CIA-RDP83M00171 ROO1800110002-5
analysis in CIA. The first session discussed the topic
in general, while the second and most recent seminar was
devoted to a specific problem: the usefulness of multi-
disciplinary analysis and the difficulty of recruiting or
training analysts for this sort of research. In subse-
quent meetings, the participants will take up other
particular features of the state of analytical work.
Whether that work can be appreciably improved by the use
of particular new methods has not yet been considered at
any length, but such consideration is in any case less
important than a more general examination, which such
seminars encourage, of the pertinence of the assumptions on
which traditional forms of research are based.
SECRET
Approved For Release 2003/O9/(?Y: CIA-RDP83M00171 R001800110002-5
 Approved For Release 20031Q c f \ -1RDP83M00171 R001800110002-5
The Defense Intelligence Agency
Directorate of Intelligence
The Directorate of Intelligence uses several mathe-
matical models in its analysis of the forces of the Soviet
Union and the Warsaw Pact. The Navy Branch of the Soviet/
Warsaw Pact Area Division has, for example, recently begun
to use a model that aids in judging the effectiveness of
Soviet ballistic missile submarines. The model may eventu-
ally be used in analyses of other types of submarines and
ships.
Another office in this division, the Military Geography
and Movement Branch, is using a computer simulation model,
DIA
25X1
k study the ability of the Warsaw
Pact to transport supplies. The model can be applied to
any transportation network for which a study of the move-
ment of, for instance, munitions is sought. Analysts in
DIA are now combining features of a file of logistical
DIA
25X1 data with
and are considering a similar
connection of the model with a file on ground order of
battle. Such improvements would permit more detailed
analysis of the ability of the Warsaw Pact to move men as
well as supplies.
The Physical Vulnerability Branch of the Installation
Intelligence Division has adopted techniques used in engineer-
ing to the development of intelligence about targets. One,
Approved For Release 2003/ cR RDP83M00171 R001800110002-5
29
 25X1 DIA
Approved For Release 20 Mgff A-RDP83M00171 R001800110002-5
Directorate of Estimates
The Long Range Forecasting Division of this Directorate
has among its responsibilities that of encouraging the
application of new analytical methods. To this end, the
Division answers specific requests for help from analysts
in the Directorate of Estimates, primarily by demonstrating
the methods, statistical programs, and files that can,
without unacceptable costs or delay, be applied to the
analysts' concerns. In addition, an officer in the Division
offers informal instruction in various new methods for those
whose work would benefit from their use and regularly
informs analysts in the Directorate of the methods and
programs that might be applied to their research. The
Division finds that the constant press of this kind of
assistance precludes experimenting with more speculative
techniques.
The Directorate is presently using a number of
analytical methods and programs, developed in some cases
SECRET
Approved For Release 2003/091L~ : CIA-RDP83M00171 R001800110002-5
 Approved For Release 2005 RECTA-RDP83M00171 R001800110002-5
with the assistance of contractors, in some cases independ-
ently. Most are designed, of course, for the analysis of
military and particularly strategic problems.
1 LcT
Approved For Release 200 CIA-RDP83M00171 R001800110002-5
31
DI
25X1
DI6IA
 Approved For Release 200-9 ON R-RDP83M00171 R001800110002-5
DIA
25X1
Several arsenal exchange models are also in use in
the Directorate of Estimates.
Also used in the Directorate of Estimates is a
specialized model of arsenal exchanges designed by CIA,
which employs the model as well.
The Support Center of the National Military Command
System has designed a model for estimating the immediate
effects--fallout and blast--of nuclear weapons. This, the
has been used
extensively in war games, and in the preparation of Defense
Intelligence Estimates.
DIA
25X1
DIA
25X1
Approved For Release 2003/69 DP83M00171 R001800110002-5
 Approved For Release 200 1 R:ECTA-RDP83M00171 R001800110002-5
Directorate of Collection
The Collection Analysis Support System, a product of
working on a contract for DIA, is a
set of statistical computer programs that has been used
by the Directorate of Collection to gauge the costs and
effectiveness of various systems for collecting intelligence.
These programs include several for the portrayal of statis-
tical information on maps, a geographical file derived
from CIA's World Data Bank I, and a statistical package
for use in various social sciences. I
While the Collection Analysis Support System as so
far been used primarily in studies of collection systems,
it has also been experimentally applied in the actual
analysis of intelligence. It has been successfully
employed, for example, in the preparation of ground order
of battle, and analysts in DIA are considering other uses
of this sort. '
SECRET
Approved For Release 2003/09/2 : CIA-RDP83M00171 R001800110002-5
 Approved For Release 2003/fJ RDP83M00171 R001800110002-5
Defense Intelligence School
During both the Management Course and the longer Post-
graduate Intelligence Course of the Defense Intelligence
School, students are introduced to features of the DIAOLS
computer system and to the use of certain quantitative
techniques, such as Bayesian analysis and queueing theory.
Decision tree analysis and the program evaluation and review
technique are also discussed. Officers from CIA's Informa-
tion Science Center provide lectures on these methods, and
the instruction is supplemented with a text prepared by the
American Management Association. While the survey of these
techniques is designed primarily to educate students in
their use in management, it also acquaints them with the
ways in which they can be employed in the analysis of intel-
ligence.
ECRET
Approved For Release 2003 9W : CIA-RDP83M00171 R001800110002-5
 Approved For Release 2003*9R?i4-RDP83M00171 R001800110002-5
The Defense Advanced Research Projects Agency
The Advanced Research Projects Agency supports basic
and applied research in a variety of new analytical methods,
some of which have already been employed by the intelligence
community. ARPA is paying for or conducting studies of
decision theory, quantitative indicators and forecasting
techniques, methods applicable to early warning, and arti-
ficial intelligence.
ARPA is funding six major contractors to conduct
studies in the field of decision theory which analysts at
ARPA believe has a number of very promising applications
to the work of some intelligence analysts. It is expected--
and has in some instances been demonstrated--that decision
theory can be applied to problems in forecasting, the ex-
pression of uncertainty in estimates, the study of negotia-
tions, and the allocation of resources. With respect to
forecasting, decision theory should encourage, not the
elimination of subjective judgments, but more rigorous
and explicit expressions of them.
The development of quantitative indicators and of
forecasting models and techniques has been another concern
of ARPA's research. In one program, ARPA, with the as-
sistance of Consolidated Analysis Center, Inc., is working
on numerical indicators of international and domestic
events. These indicators suggest trends in and relation-
ships among various political, economic and military
SECRET
Approved For Release 2003/09 : CIA-RDP83M00171 R001800110002-5
 Approved For Release 20e[ / ffA-RDP83M00171 R001800110002-5
events and will, it is hoped, provide analysts a novel
and useful means of interpreting and forecasting events
within and among countries. In addition, models for fore-
casting distant and more immediate political and military
environments have been devised and are now being applied to
different regions of the world.
ARPA has also assisted in the development of complementary
political indicators for the same purpose, but such indi-
cators have not yet been completed or tested.
The Interactive Analyst Station, a project of ARPA
and the Rome Air Development Center, is an attempt to
develop advanced methods for indications and warning
analysis. In this project, the Synectics Corporation,
the Cambridge Project, and the Center for Computer-Based
Behavioral Studies at the University of California at
Los Angeles have developed a variety of techniques which
employ computers and which facilitate both the retrieval
and the analysis of information. ARPA is studying these
techniques, assessing the degree to which they support
indications and warning analysis, and attempting to apply
SECRET
Approved For Release 2003/092 : CIA-RDP83M00171 R001800110002-5
 Approved For Release 200 Ei-RDP83M00171 R001 800110002-5
to such analysis those techniques that have promise. A
preliminary system has been developed which can be used
for demonstrations and tests. In the next phase of the
program this experimental system will be refined and then
further tested at the Strategic Air Command.
ARPA has also sponsored a program to study the security
implications of the exchange of technology and of expanded
trade with the USSR, Eastern Europe, and the People's
Republic of China. The elaboration of models of the proces-
ses by which the USSR, the PRC, and Eastern European countries
apply technology obtained through trade and exchange agree-
ments has been undertaken. New information and the results
of certain case studies will be used to test these descrip-
tive models and will provide a basis for judging their use in
prediction. The results should be helpful in determining
the strategic implications of the transfer of technology.
The study may also provide the understanding needed to
make more critical judgments on the direction that Soviet
military technology is taking.
Several offices at ARPA are investigating the possible
uses of artificial intelligence. The latter is not a
single technique but a relatively new branch of computer
science that draws upon a number of disciplines--psychology,
logic, epistemology and engineering, among others--in an
effort to understand certain intellectual processes and,
when it is possible, reproduce them with a computer. There
SECRET
Approved For Release 2003/091 : CIA-RDP83M00171 R001800110002-5
 Approved For Release 20SAEr-RDP83M00171 R001800110002-5
are a number of ways in which such an approach might be
applied to intelligence. It may be possible, for example,
to prepare a model of the ways in which decisions are reached
in a foreign government, but a model distinguished from
others by the addition to it of the ready, intuitive
judgments that analysts have formed concerning politics
in that country. In one such application, ARPA's Tactical
Technologies Office is supporting the development at the
Rand Corporation of a model of terrorism. The work of
ARPA's Information Processing Technologies Office on the
manipulation of very large data bases and on "intelligent
terminals" also makes use of artificial intelligence. (Intel-
ligent terminals are computer terminals that will permit
analysts to do more quickly such routine tasks as searching
their files, keeping records, etc.)
SECRET
Approved For Release 2003/093/02 8 : CIA-RDP83M00171 R001800110002-5
 Approved For Release 200339/ R -RDP83M00l71 R001800110002-5
The State Department
The State Department's Bureau of Intelligence and
Research has in the last few years commissioned several
studies directed, in part, at determining the usefulness
of quantitative techniques for research in foreign affairs.
The most ambitious, Project Quest, was conducted by
PRINCE Analysis, Inc., under a contract with the Bureau's
Office of External Research. The study attempts to offer
a systematic answer to the question of the usefulness of
new analytical methods.
The authors wished to discover whether it was possible
and desirable to use in the work of the Department of State
the frequently quantitative methods employed by academic
students of comparative politics and international affairs.
To do so they undertook first to examine the work of
analysts in INR from the perspective of a social scientist.
They concluded that the analysts use quantitative techniques
infrequently and in thoroughly traditional ways; that the
analysts make predictions that are concerned with the im-
mediate future or which fail to specify the period to which
they apply; and that the analysts' concerns are considerably
broader and more complex than those of their academic counter-
parts. This last observation was especially pertinent, for,
according to the study, "in spite of more frequent and sus-
tained attempts to deal with these analytical complexities,
SECRET
Approved For Release 2003/09M : CIA-RDP83M00l71 R001800110002-5
 Approved For Release 5R(F1: CIA-RDP83M00171 R001800110002-5
the quantitative scholar in comparative politics and inter-
national relations has not made sufficient progress at this
time to be of use to the foreign affairs analyst."
While existing quantitative research thus appeared
to offer little assistance to the considerably more demand-
ing work of INR's analysts, it still. seemed possible that
particular quantitative techniques might bring greater rigor
and precision to bear on specific analytical problems. The
authors devised six case studies to see whether this was so,
choosing topics which were pertinent to the work of INR's
analysts and which lent themselves to treatment by one or
more quantitative methods. (Among the techniques used in
the studies were regression analysis, correlation, and
computer simulation and modeling.)
The case studies suggested that certain techniques
could assist analysts in some but by no means all aspects
of their work. For what the authors called "information
gathering," quantitative methods, such as content analysis,
were found to be useful: they encouraged clarity, the
recognition and analysis of trends, and the systematic
comparison of events. Employing such methods also made
easier the use of information already in numerical form,
like data on international trade, voting and the cost of
living.
The quantitative methods used in the case studies
were less useful for explaining events--for "testing
Approved For Release 2003/SfCR 1RDP83M00171 R001800110002-5
 Approved For Release 204 FDQRETIA-RDP83M00171 R001800110002-5
assumptions and hypotheses"--primarily because of the great
variety of influences that an analyst must consider. The
authors concluded that the final activity into which they
divided the work of an analyst, forecasting, might at times
benefit from the use of such methods as correlation tech-
niques, trend analysis, and the Delphi technique.
INR has commissioned other studies similarly aimed not
at applying an analytical method to a particular issue but
at determining more generally whether new techniques will
further the Bureau's work and, if so, how they should be
used. A paper written for INR by Lincoln Bloomfield of
MIT's Center for International Studies reviewed the recent
literature of the social sciences, particularly that con-
cerning the development of theoretical structures to guide
analysis, for the value of this work to the improvement of
the policy process. Bloomfield's survey led him to a
number of suggestions for applying developments in the
social sciences to the work of the State Department. He
proposed that in their reporting, Foreign Service officers
keep in mind explicit theories of political behavior, such
as the Bureaucratic Politics Model. He recommended also
that analysts be trained in the use of such formal fore-
casting methods as the Delphi technique. The ability of
the State Department to anticipate issues would be improved,
he thought, by the readier acceptance of dissent within
SECRET
Approved For Release 2003/09/40? : CIA-RDP83M00171 R001800110002-5
 Approved For Release 2O EtG lIA-RDP83M00171 R001800110002-5
the Department, by the use of political games, and by a
continual, critical examination of the policy process.
He urged that the Department undertake case studies of its
own operations and that periods of routine activity as well
as operations during crises be studied.
The Center for International Studies also conducted
for the State Department a study of "global interdepend-
ence"--the ever closer economic, cultural, political, and
strategic ties between states. As part of the project,
members of the Center examined in some detail the question
of whether it was possible to devise methods and concepts
that would aid the study of interdependence. They con-
cluded that there were indeed certain "methodological
maxims" that should be borne in mind. The maxims specify
the ways in which analysts should consider such influences
on the interdependence of states as the international
milieu; the policies, resources, and histories of the
nations under study; etc. (Among the pithier maxims was
this helpful injunction: "Sequentially explicit construction
heuristics [design principles] must be explicated; other-
wise their operational meaning remains obscure.")
R ET
Approved For Release 2003/D9/02 : CIA-RDP83M00171 R001800110002-5
42
 pproved GINTELLLIIGENCE/COMMUNITY 0J7o1 Q018001100 t-i
AD/ DCI/IC
D/DCI/IC
This is a comprehensive survey by
I I of the Community's use of
new analytical methods. It is the successor
to an earlier study done on contract but never
published because of certain weaknesses. I
think this is a good job, worth disseminating
to USIB agencies and perhaps to the Working
Group of the NSCIC as well.
C/PRD
Approved For Release/~Tt@ft00171 R0018001100
 rov-PyMm 2O /09462149IC-RDP83M00171 R001800110
c a~ ~0 5 S h^
a6~^
CdYr~ (DATE 1
a /09702: M08u1Qa 1
04prd FA
FORM
N
O 10-
AUG
54 101 WHICH MAY F BEMUSEDiDi
 p
provc"oraRel e f 00 / 0 1 R00180011000
VKAA
pproved For Release 2003/09/02 : CIA-RDP83MOOl71 R0018001100
I
 pproved F }~~~ La CIA-RDP83MOQ171 RO01800110002
9NTELL~(938Y06 NOMMUNITY STAFF
8 January 1976
NOTE TO EO/ICS t~,
1. An evaluation of the methods that are
described in this survey would be useful, but
we believe an attempt at one now might be
premature. In many cases the analysts who
are trying to bend these new techniques to
their own purposes are themselves uncertain
how productive their efforts will ultimately be.
Until such judgments can confidently be made,
moreover, it will be difficult to determine
which of several approaches to this still
experimental work will prove most rewarding.
(The paper does offer some tentative judgments
on this question. )
2. Despite the (considered) omission
from the paper of an evaluation of these
methods, it should nevertheless be useful
as a survey of the Community's work in this
field, and we recommend that as such it be
forwarded to Mr. Knoche and Gen. Wilson.
PRD
Attachment:
A Survey of the Community's Use of
New Analytical Methods
Approved For Release 2003T:10 - DP83M00171 R001800110
 Approved For Release 2003/09/02 : CIA-RDP83M00171 R001800110002-5
Approved For Release 2003/09/02 : CIA-RDP83M00171 R001800110002-5
 1n mnv1975
MNNI uvGU rut r~cicapc 9~c~ri ~r~- tCV'Ui 11V P r~u~ vvi iuuuc-J
can.p~~d c
Approved For Release 2003/09/02 : CIA-RDP83M00171 R001800110002-5
COMMENTS: