Schneier on Security
A blog covering security and security technology.
« Friday Squid Blogging: Squid Embryos |
| John Mueller on Nuclear Disarmament »
June 22, 2009
Engineers More Likely to Become Muslim Terrorists
Time to start profiling.
EDITED TO ADD: here's the paper.
Posted on June 22, 2009 at 7:10 AM
• 39 Comments
To receive these entries once a month by e-mail, sign up for the Crypto-Gram Newsletter.
As a Muslim Computer Scientist, I wonder where I fit in...
So we can further profile on "left handed" (sinister, gauche, cack handed) and those on the "Autisum spectrum" then...
I just get the deep down feeling that it's somebody who does not like "beards making" it up as facial hair is more prevelent in engineers...
from the article:
"and jihadist recruitment manuals focus on a personality profile rather than technical skills."
If that is the case, you'd think profiling should be more effective in finding them.
That was an interesting article and the end very nicely clarifies by saying "the probability of a Muslim engineer becoming a violent Islamist is minuscule."
What the *bip* is this piece of *bip* of statistical analysis?
"In 16 other countries we investigated, engineers seem to be no more right-wing or religious than the rest of the population, but the number of engineers combining both traits is unusually high."
That makes no sense whatsoever.
"dislike of democratic politics which involves compromise" -> they don't seem to know what is the job of an engineer. If we are not discussing and making compromises all the time, I do not know what we're doing.
Basically, almost all third-world people taking on higher education will go to engineering, not "british poetry". So I guess engineers are under-represented, not over-. But the real point is that educated people are more likely to become terrorists than, say, poor farmers, even so they suffer less, are less oppressed, and should know better about the outer world.
> "In 16 other countries we investigated,
> engineers seem to be no more right-wing or
> religious than the rest of the population, but the
> number of engineers combining both traits is
> unusually high."
> That makes no sense whatsoever.
Sure it does. It just means they are not statistically independent.
Suppose 50% of the population is rightwing, 50% is religious, than you'd expect 50%*50%=25% to be both, assuming they're independent variables. But if you find that, say, 40% is both, than that is unusually high.
Of course I wouldn't expect these traits to be really independent. But apparently, a significantly higher percentage of engineers has both than is usual in the general population.
Yeah, but when I read the article, they do not say anywhere "compared to the percentage in the general population".
Ill effects of education :-)
Seriously .. is it surprising? Only those with some background in science will know how to hook up vibrator on cell phone to a trigger a bomb - you don't expect english (or arabic) majors to do it.
In fact, we see that certain variations of Christianity go with certain political stands in the US. If I describe somebody's politics, you can make a fairly good guess at their political views (obviously you're going to be way wrong a lot), and vice versa.
Now, I don't think polls like this address religion in a sense I'm comfortable with, so I always want to know what signifiers they're using for religion. "Churchgoing" is nice and measurable. "Religious" is not.
(About the article's authors... and their book deal...)
Diego Gambetta is official fellow at Nuffield College, Oxford. Steffen Hertog is Kuwait Professor at the Chaire Moyen Orient-Méditerranée, Sciences Po, Paris. Their book, Engineers of Jihad, will be published next year by Princeton University Press
That there are more engineer terrorists than fine arts terrorists is hardly surprising: there are more engineers, period.
The statistics in the article would be more illuminating (to a critical reader) if the figures were somehow normalized for the proportion of engineers in the population at large.
Of course, expressing statistics in a meaningful way just makes Joe Average's eyes glaze over.
So, it would seem that most terrorists are engineers...so sad that that doesn't say anything about which engineers will likely become terrorists.
That is, we've found a correlation, but the finding is retrospective, not predictive.
I expect that there is something which makes prospective terrorists become engineers, rather than something which makes engineers become terrorists.
The statistics of this article looks rather suspect to me.
Of the identified ‘terrorists’ with higher education, 78 out of 178 had studied engineering, 34 Islamic studies, 14 medicine, and so on.
However, these figures do not look to take account of the total numbers of people (and particularly those from Islamic countries) that study these subjects. Therefore we do not see the relative proportions of engineers etc, but only the total numbers.
A quick look at the website of the UK’s Higher Education Statistics Agency gives the following for 2007/08: http://www.hesa.ac.uk/dox/dataTables/...
This shows 139,435 engineering students (with 30,965 being non-EU), compared to 14,510 theology students (1,175 non-EU) and 61,810 medicine and dentistry students (6,515 non-EU). If similar figures are representative of those from Islamic countries, irrespective of where they studied, we would expect any grouping of higher educated persons to contain very approximately 9.6 times as many engineering graduates as theology graduates (actually 2.3 times in the ‘terrorist’ group). Also we would expect to find very approximately 2.3 times as many engineering graduates as those with medicine or dentistry (actually 5.6 times in the ‘terrorist’ group), and very approximately 0.2 times as many theology graduates as with medicine and dentistry (actually 2.4 times in the ‘terrorist’ group).
In so far as the referenced article’s terrorist tendency figures mean anything, they show a very high representation for those who did theological (including Islamist) studies: over 4 times the representation of engineering graduates and over 10 times the representation of medical and dentistry graduates.
That is unless university theology studies are much more prevalent in those originating in Islamic countries. [Note: the figures are also somewhat confused by the detailed definitions of the subject studied.]
The reports claiming some kind of corellation between engineering studies and terrorism are not all that new.
Over a year ago, Oxford University published a sociology paper 2007-10 "Engineers of Jihad" by Diego Gambetta & Steffen Hertog.
A Design News story from April 2008 gives an overview of the paper and its findings, including the authors believing that the results were often misunderstood.
Terrorists are an extremely small part of the general population that tries to avoid attention before becoming active. In effect, a terrorist that has not gone active is only a potential terrorist.
These two things make them essentially impossible to profile: There is just nothing to go on.
Ordinary profiling starts with a crime and then tries to find out what person commited that crime (or series of crimes). This way the individual characteristics of the perpetrator can be deduced. Every criminal is different.
As a result, you either get a group of potential terrorists so large that it is useless. Or you go specific enought with the profile and have a good chance of missing all the terrorists.
Come to think of, this is what makes terrorism effective: You cannot really prevent it, except by not pissing people off enough to commit it.
Good catch webbnh and Nigel Sedgwick. It is still an interesting article but only because I wouldn't have guessed engineering was such a prevalent major.
As a fat old white guy injineer, sure go ahead profile me. I'm already being profiled: for having supported Ron Paul; being anti abortion; buying gold coins; being Catholic; being pro-life; belonging to the NRA and GOA; owning guns; speeding; being revolted by taxes; disgusted with politician of both alleged sides of the aisle; and countless other aspects of my beliefs.
So what's one more bull's eye painted on my back!
Yes, the fact that you are mostly in a comfortable majority yet think you are persecuted is definitely a characteristic of an irrational profile. The short version of what you're describing is called "victim mentality". Probably should seek professional help and learn to be happy with your success instead of angry at others.
If I were a terrorist, the first thing I would do is take a short vacation. If I was allowed to fly round trip without incident, I would know I'm not on the government's radar.
A pretty handy-dandy service for the other side, if you ask me.
The author is bending over backwards near the end of the article to avoid saying "autism" or "Asperger syndrome". Interesting speculation, but obviously more research is needed.
In the meantime, this is another great reason why we programmers need to tell more people that "software engineering" isn't actually real engineering! (Also, for the benefit of any profilers reading this, I have an established record as a tree-hugging hippie liberal.)
"I wouldn't have guessed engineering was such a prevalent major."
It depends what you mean by engineering...
The definition is somewhat skewed by engineering departments taking on "ascociated subjects" to keep the numbers up.
For instance computer programing (which many will tell you is not an engineering subject but an art) often falls under either the maths or engineering departments depending on the type of university (old/new) and type of course (pure/applied).
Then there are other subjects such as architecture which again may be refered to as engineering, various primarily chemical, biological, and physics courses have engineering in their title and course material.
I tend to view it this way these days for the first and second worlds,
You have scientists who earn a crust by beging money to do research, engineers who take the work of scientists and apply it to systems and technicians who take the work of engineers and build the systems and maintain them. With very broad overlap in each area.
Oh and then you have the other 80% of the people who need those systems to maintain their living one way or another.
Unfortunatly due to the way some people think, people get given "job titles" instead of money.
So in the UK where the use of the terms engineer and technician are unregulated you have people that change the tires and exhausts on your car calling themselves "engineers"...
Ah! but you forget the tens of thousands of MCSE (Microsoft Certified Systems Engineer) that is what I call a potent terrorist task force... :)
Where do they ask:
Of the general population of those seeking degrees in Germany who were from one of the 26 Islamic countries, and who were Muslim, what percentage were seeking Engineering degrees? Does that figure correlate to the percentage of Engineering degree seekers among the population of known terrorists?
Those seem like obvious questions to investigate, but there is no mention of them in the article.
Unless the percentage of Engineering degree seekers among known terrorists significantly higher, then the question of "why Engineers?" becomes pretty uninteresting. The answer "because given the economy and society of the degree seeker, Engineering is the surest pursuit to guarantee a good living" seems good enough.
Going up in a balloon over the pop-psychology of motivation in pursuing various degrees might be entertaining in a late night dorm room discussion, but as a public discussion of security, it looks myopic in ignoring general populations and taking meaning only from the few that have grabbed their interest.
It looks like the authors do make a comparison to the percentage of engineers in the population in their paper.
I don't know why that was left out of the New Scientist writeup, since it's obviously essential.
I've removed one comment. Do not use this thread as an excuse for religion-bashing.
Meh, I would suspect that people's reasons for becoming terrorists are highly personal.
Having said that, profiling individuals on broad criteria probably isn't going to get anyone anywhere. It seems more the sociological approach where it would seem the psychological approach might yield some more results...
None the less, its not like terrorists are gonna feel obligated to talk about their feelings *after* they were caught.
I think its been echoed, but is this method EVER going to produce a pratical solution to anything? I think it just finds another direction for overly-paranoid people in high positions to point their fingers/blame and make it look as if they are doing something good for the common folk.
I don't find it surprising at all -- folks like "reinkefj" are actually very well over-represented among engineers as opposed to scientists. I doubt it's an Asperger's type thing -- they generally don't show traits of inability to understand what others are thinking.
But general "crankness" is a common feature of engineers -- with of course the extreme "crank" position being terrorism and cultish political positions. I'd expect it to be more tied to being extremely trained in one small area to the exclusion of almost anything. The high education leads one to be trained to deeply consider most issues; but the ignorance of other fields leads one to make errors in their analysis. For example, if you apply a rigorous analysis of politics, as an engineer would, while lacking a wide education in history, you will almost inevitably lead you to serious errors of correlation vs. causation -- lack of enough data is dangerous.
And of course, paranoia is then common, since the correlations that are mistaken as causation will create an abstraction with "hidden hands".
This is an ethnological problem, and not really a statistical one. When you get into this region of anthropology, statistical analyses are more often than not misleading. Systems of meaning rarely follow any of the usual statistical distributions; without a good theoretical reason for a given distribution, you're just making crap up by presenting it statistically.
Surely you can't expect to confidently blow things up without at least a little engineer training?
Well, Bruce, I've subscribed to New Scientist since 1976, but this piece, last week, was one of the best examples of non-science I've read for a very long time. I won't even start to address their 'grasp' of statistics.
However, after 23 years in security, I can state unequivocably that almost all acts of terrorism are false flag operations. So I wonder by what criteria the authors compiled their sample of 'terrorists'?
Engineers, usually electrical, are over-represented among germ theory deniers, HIV deniers, creationists, anti-relativists, and the various other pseudoscience nuthouses.
At Uni engineers don't get invited to any "groovy" parties. After social isolation, doomed to a life of being rejected by the opposite sex, and being involved in a career that many don't respect, some adopt dangerous attention seeking behaviours.
Well it's a theory.
Solution - be nicer to engineers.
"as greater intolerance of ambiguity, a belief that society can be made to work like clockwork, and dislike of democratic politics which involves compromise, are more common among engineers."
The discipline problems (pr0n surfing mostly) we've had have nearly always been the engineers.
I've put it down to arrogance. An over valuing of their own intelligence and lack of valuing what their manager, customer or larger system requires. The only control I could think of was strong doses of ethics training (they wouldn't let me administer emetics)
Thanks for posting the link to the paper. That was very interesting.
Whatever happened to probable cause?
@Clive Robinson: "It depends what you mean by engineering..."
I was actually also wondering about that issue but it was covered in the paper:
"Engineering in our definition includes computer sciences and architecture. The latter was included because it is commonly part of engineering faculties in Middle Eastern countries, as it is in European countries."
An engineer who helped designed submarines for the Australian navy once tried to convince all the guests at a party that I was mentally ill because a "confidence trickster" had sold me "carved rocks" that I, poor deluded soul, thought were dinosaur bones.
At which point I remembered that most "scientific creationists" are engineers, not biologists (and most Christians who are biologists believe in evolution), and Australia did have a lot of trouble getting their home made submarines to work properly.
While I only skimmed through the paper, I can't help but recall that Mathematics and Engineering were always highly valued in Arab cultures, before and after the rise of Islam. I had to chuckle right out of the gate at "modern Islamic radicalism". Apparently disco is what separates the modern from the past. I liked the non technical folks (includes the authors I presume) being labeled as latecomers to join a movement. I wonder if that carries over into general behavior on actionable agendas other than "Islamist mobilization". One just has to wonder, are engineers prone to extremism or simply prone to try and get things done.
Now if I were an evil engineer in one of these countries constantly blowing itself up, I could see backing a group that blew stuff up so I would always have work rebuilding it. Ooh, maybe I should dig up some numbers and publish about that...
Hey, guys! I've got a really good idea! Let's use a piece of psuedo-science as an excuse to bash other people's majors! Because we all know those silly EE/ME/Math/Basketweaving majors _all_ fit profile X, and therefore I have to assert how glad I am that I'm not a EE/ME/Math/Basketweaving major. Seriously, grow up.
-An *Atheist*, *Left-wing* Engineering/Mathematics student
Schneier.com is a personal website. Opinions expressed are not necessarily those of BT.