Growth of mujahidin in Afghanistan and Baysian belief networ

[Humera Khan]

Hello,

Just to jump into the fray, it would seem that the why and the how are
not being clearly separated:

1) terrorists commit acts of terror because they "work".

The fact that some of the attacks "work" does not address why they
happen. Almost all terrorist groups claim that they are fighting
unjustice, real or perceived. They are really not doing it just because
they can get away with it- they always have some cause that they feel
very strongly about.

I really doubt terrorists would have the need to use any scientific
method to recruit when they can simply raise sympathy by loudly
proclaiming "injustice".

2) Terrorists tend to claim that if the system in place does not help
them, and instead furthers their oppression, they will turn to scare
tactics. They try to break the system in the hopes it would help them.

Given the desperation of these people, and the way humans tend to act,
it is virtually impossible to prevent all the modes of attack that can
happen.

I think one of the biggest failures of intl policy has been that govts
fail to isolate WHY terrorists feel the need to lash out against
civilians, and instead only focus on HOW to prevent the action. Or of
course, they isolate the why but it is not economically in their best
interests to pursue a policy that addresses the problems.

Personally I think a public study that focuses on the need to examine
and address the why of terrorist attacks rather than just on how to
prevent an attack would be more meaningful in the long run.

Sorry if this is begininng to sound like an SD 101 class "look at the
cause, dont just treat the symptoms!"

-Humera
Humera Khan <salsabeelhk@yahoo.com>

[Justin Lyon]

Hi everyone,

Dawnellen and I have been working on using SD to
analyze the growth of mujahidin in Afghanistan from
the 1970s to the early 1990s (working in
collaboration with colleagues at the UN and under the
guidance of Dr. Warren at London Business School via
the MIT advanced study program in System Dynamics).

Similar to the work done by HPS, we hypothesize that
the accumulation of anger reaches a threshold at which
point people are motivated to take action -- either
flee and become refugees or perhaps join a rebel group
(these are only representative of possible actions).
That is, anger affects the "Rebel Recruitment Rate" --
the flow from non-rebels into rebels.

Of course, getting any sort of real data on the
accumulation of anger in a given population is
difficult (especially in areas like Afghanistan), so
weve started working with people in the field of
political psychology. This field has developed methods
of gathering quantitative data relating to
psychological issues in these types of situations.

In one discussion, a politcial psychologist brought up
the fact that he is using . . . "Baysian-based belief
networks to represent causal logic chains and
cumulative probabilities in complex decision
environments" . . . "to create dynamic models of
social - psychological processes." The software they
are working with is called NETICA.

QUESTION
I know zip-all about NETICA or Baysian-based belief
networks.

How does this approach compare with system dynamics
approaches? Are they complementary?

Id love to share with him any research in the field
that compares or contrasts his approach with System
Dynamics (similar to the paper I read written by Hans
Scholl at Albany on "Agent-based and System Dynamics
Modeling: A call for cross-study and joint research")

Does anyone know of this kind of research?

Thanks,
Justin

["geoff coyle"]

Folks,

This correspondence about the Rev Dr Thomas Bayes, the Mujahidin, and
Afghanistan reminds me of the USS Maine (bear with me, there will be two
points).

Point 1 (a little light-hearted). The sinking of this battleship in Havana
harbour, apparently as a result of a terrorist act, led to the
Spanish-American War, fought on a battle cry Remember the Maine!. Correct
me if Im wrong, but I believe that the subsequent enquiry showed that USS
Maine had sunk as a result of an accidental explosion on board. However,
Remember the Maine! was a case study in one of the earliest modern texts
on Bayes. I wonder if General Shafter was acquainted with Bayes. Shafter,
who commanded the expedition to Cuba (assisted, or supplanted, by Teddy
Roosevelt, who invented national parks and Teddy bears), suffered from the
handicap that his gut hung down below his thighs when he sat down to write
his orders. Even Bayes is a simplification of reality.

Point 2 (and more serious). The Mujahidin were not terrorists in the sense
we now use. They were the freedom fighters
who opposed the Russian invasion of Afghanistan. The Taliban came later -
the word is the plural of Talib, literally meaning a religious student,
though perhaps not one the Anglican Church would recognise as such. Al Qaeda
are another lot, often not
citizens of Afghanistan (ethnically and linguistically, there is really no
such person as an Afghan), and it is they who attacked New York. Its
important to discover the issue history (or to find out some facts) before
starting to model. Hypothesising about hatred (by whom and of whom or of
what and measured in what way?) leading to recruits may be a shade on the
simple side. A powerful person who hates may be able to conscript, making
recruitment irrelevant.

Point 3 (OK, I lied about two points) is to reiterate my plea NOT to do this
sort of work over the public air-waves. If you can get interest from someone
responsible (the FBI, CIA, or Uncle Tom Cobley), by all means do it in
private.

Geoff Coyle
From: "geoff coyle" <geoff.coyle@btinternet.com>

["Weaver, Elise A"]

Bayesian belief updating is a statistical technique for integrating
recent evidence with prior beliefs. It is at odds with classical
statistics, where you start from scratch with each new experiment.

In your "anger" context, I would assume the application is as follows:
If you start with the prior belief that someone is your enemy, you
will require a lot of evidence to the contrary in order to change
your belief.

From: "Weaver, Elise A" <eweaver@WPI.EDU>

["Raymond T. Joseph"]

Geoff,

<snip>
Point 3 (OK, I lied about two points) is to reiterate my plea NOT to do this
sort of work over the public air-waves.


Please elaborate on the purpose of a keeping such a study (or any study)
from the eyes of the public. I ask this as I am a firm believer that the
truth will set you free.

Ray
From: "Raymond T. Joseph" <rtjoseph@ev1.net>

["Bob Walker"]

Raymond;

Geoff will undoubtedly reply to you as well but how about this for starters;

Can you imagine a Poker Game where all hands are dealt "face up" on the
table? Not much fun.

So, lets modify this... now only selected players, lets call them "the
good guys" have all of their cards exposed to the world... there is a bet
after each round... which players are likely to raise the stakes?

Bob

Robert J. Walker
System Dynamics Practice Leader
Delsys Research Group, Inc.
From: "Bob Walker" <walker@delsysresearch.com>

["Lazaros Petrides"]

> Please elaborate on the purpose of a keeping such a study (or any study)
> from the eyes of the public. I ask this as I am a firm believer that the
> truth will set you free.


I understand that it was Geoff Coyles view that were asked for by Raymond,
but since he sent the message to the list I would assume that anyone could
reply.

Given that the answer seems obvious to me, and fearing that it is so obvious
that Professor Coyle may not find it necessary to reply, I will try to
provide a specific answer:

Assume that one of the result of the study is that terrorism occurs because
it actually works. Whom would the general knowledge of this conclusion set
free? Could it be that terrorists would find it easier to recruit more
people since theyve got evidence (not only empirical which can be cooked
and argued against, but also theoretical and therefore scientific) that
their methods work?

This analysis could obviously be lengthier

Lazaros
From: "Lazaros Petrides" <lpetrides@cwcom.net>