The statistics of fear

Posted on by

My mom is always worried about her two sons who live on the opposite side of the nation from her. Before cell phones were common I would return home to find messages like the following on my answering machine (I am paraphrasing):

“Abhi, beta. Please tell me that you aren’t eating beef. You will get Mad Cow disease, you will become a vegetable, then you will die.”

After hearing this message on half a dozen occasions I pretty much gave up beef. Why? Because this is how the messages usually ended:

“Promise me beta, ok, love mom, bye.”

Or what about this more recent one on my cell phone:

“Abhi, I heard there are fires all around Los Angeles, be careful, stay away from the hills.” [note: I am nowhere near the hills]

My favorite to date has been:

“Abhi, do you drink water out of plastic bottles that you re-fill? The plastic leaks chemicals into the water. You will die.”

Since the World Trade Center attacks and the terrorist attacks that have followed in other parts of the world (like the recent Mumbai Train attacks), many people have established a new dichotomy in their minds. There was before 9/11 and there is after 9/11. “Everything is different now.” I find such sentiments bordering on delusional but until now I have had no really substantial counter-argument to point to that was any more cogent than me calling the person an “idiot” . That changed this week when John Mueller of Ohio State University published this paper for the Libertarian Cato Institute. Titled, “A False Sense of Insecurity?“ the paper takes a look at how ignorance of statistics allows entities (or my mother) to use fear inappropriately. This article (only five pages) is a must read and something I wish every American was exposed to.

For all the attention it evokes, terrorism actually causes rather little damage and the likelihood that any individual will become a victim in most places is microscopic. Those adept at hyperbole like to proclaim that we live in “the age of terror.” However, while obviously deeply tragic for those directly involved, the number of people worldwide who die as a result of international terrorism is generally only a few hundred a year, tiny compared to the numbers who die in most civil wars or from automobile accidents. In fact, in almost all years, the total number of people worldwide who die at the hands of international terrorists anywhere in the world is not much more than the number who drown in bathtubs in the United States.

Until 2001, far fewer Americans were killed in any grouping of years by all forms of international terrorism than were killed by lightning, and almost none of those terrorist deaths occurred within the United States itself. Even with the September 11 attacks included in the count, the number of Americans killed by international terrorism since the late 1960s (which is when the State Department began counting) is about the same as the number of Americans killed over the same period by lightning, accident-causing deer, or severe allergic reaction to peanuts.

But what about nuclear weapons or other Doomsday scenarios? What if the terrorists get there hands on the really bad stuff that they didn’t have before?

Obviously, this condition could change if international terrorists are able to assemble sufficient weaponry or devise new tactics to kill masses of people, and if they come to do so routinely. That, of course, is the central fear. As during the Cold War, commentators are adept at spinning out elaborate doomsday and worst-case scenarios. However, although not impossible, it would take massive efforts and even more stupendous luck for terrorists regularly to visit substantial destruction upon the United States…

To this point in history, biological weapons have killed almost no one. And the notion that large numbers of people would perish if a small number of chemical weapons were to be set off is highly questionable. Although they can be hugely lethal when released in gas chambers, their effectiveness as weapons has been unimpressive. In World War I, for example, chemical weapons caused less than one percent of the total combat deaths; on average, it took a ton of gas to produce one fatality. In the conclusion to the official British history of the war, chemical weapons are relegated to a footnote that asserts that gas “made war uncomfortable…to no purpose.” A 1993 analysis by the Office of Technology Assessment finds that a terrorist would have to deliver a full ton of Sarin nerve gas perfectly and under absolutely ideal conditions over a heavily populated area to cause between 3,000 and 8,000 deaths — something that would require the near-simultaneous detonation of dozens, even hundreds, of weapons. Under slightly less ideal circumstances — if there were a moderate wind or if the sun were out, for example — the death rate would be only one-tenth as great.

From day one the government has followed a strategy that seeks to cover their liability in case of another attack. Part of the way to do this is to keep people in fear so that if something goes wrong they can be ready with the “I told you so,” instead of making more logical, statistics-based changes.

The shock and tragedy of September 11 does demand a focused and dedicated program to confront international terrorism and to attempt to prevent a repeat. But it seems sensible to suggest that part of this reaction should include an effort by politicians, officials, and the media to inform the public reasonably and realistically about the terrorist context instead of playing into the hands of terrorists by frightening the public. What is needed, as one statistician suggests, is some sort of convincing, coherent, informed, and nuanced answer to a central question: “How worried should I be?” Instead, the message the nation has received so far is, as a Homeland Security official put (or caricatured) it, “Be scared; be very, very scared — but go on with your lives.” Such messages have led many people to develop what Leif Wenar of the University of Sheffield has aptly labeled “a false sense of insecurity.”

The paper goes on to pose three questions that should be part of any probability risk assessment used to help determine the best way to defend ourselves against terrorism:

  • How much should we be willing to pay for a small reduction in probabilities that are already extremely low?
  • How much should we be willing to pay for actions that are primarily reassuring but do little to change the actual risk?
  • How can measures such as strengthening the public health system, which provide much broader benefits than those against terrorism, get the attention they deserve?

And finally the political dimension:

There is no reason to suspect that President Bush’s concern about terrorism is anything but genuine. However, his approval rating did receive the greatest boost for any president in history in September 2001, and it would be politically unnatural for him not to notice. His chief political adviser, Karl Rove, declared last year that the “war” against terrorism will be central to Bush’s reelection campaign. The Democrats, scurrying to keep up, have stumbled all over each other with plans to expend even more of the federal budget on the terrorist threat, such as it is, than President Bush.

This process is hardly new. The preoccupation of the media and of Jimmy Carter’s presidency with the hostages taken by Iran in 1979 to the exclusion of almost everything else may look foolish in retrospect, as Carter’s secretary of state, Cyrus Vance, conceded in his memoirs. But it doubtless appeared to be good politics at the time — Carter’s dismal approval rating soared when the hostages were seized. Similarly, in the 1980s the Reagan administration became fixated on a handful of American hostages held by terrorists in Lebanon. At the time, Reagan’s normally judicious secretary of state, George Shultz, was screaming that we needed desperately to blast somebody somewhere “on a moment’s notice” — even without adequate evidence — in order to avoid looking like the indecisive “Hamlet of nations.” He apparently preferred the King Lear approach. Normally, however, only lunatics and children rail at storms; sensible people invest in umbrellas and lightning rods.

Stop reading my post already. Just go read the whole paper. I can’t wait until my mom calls me tonight. Armed with the correct statistics I shall win the argument for once.

65 thoughts on “The statistics of fear”

  2. [OT] “Abhi, do you drink water out of plastic bottles that you re-fill? The plastic leaks chemicals into the water. You will die.”

    And I thought my mom was the only who said this! What about not cooking in non-stick cookware (causes cancer) or not chewing gum while playing sports (apparently someone who attempted to both at the same time ended up swallowing it, and the piece of gum fell squarely on windpipe, chocking the poor guy to death. If I’m not mistaken, in her version it was a badminton player).

    Anyway it’s good to know that there are other paranoid ma ke bete out there.[/OT]

  3. Abhi,

    This interview with Ian Lustick from UCTV pertains to this paranoia but from a poli sci perspective rather than statistics. A snippet:

    I have in the book a quote from a leading official in charge of helping the community of scientists to become involved in the War on Terror, and to change their research, and to apply for grants. I asked this person in front of an audience in 2004 to place the War on Terror on a continuum of seriousness. “If the Manhattan Project were a 10 on a scale of 1 to 10, in which the government knew the Nazis were trying to get a bomb and we wanted to give the scientific community all the resources it needed to solve that problem. That’s a 10 — a super-serious, absolutely real problem and treated seriously by the government as such. On on the other side of the continuum is a 1, where the problem said to exist doesn’t exist at all, though it’s politically super-convenient to say so — in other words, ‘wag the dog.’ In that case there is a war said to be going on when there really is no war.” So I asked, on this continuum, where is the War on Terror?” This is a person responsible for helping get people involved in it. “On a scale of 1 to 10, is it near the Manhattan Project or Wag the Dog?” He said, “I can’t answer that in public.” I asked him in private and he said, “Somewhere between 1 and 1.5.” So, it’s not as if inside the government what I’m saying is not known, it is just that no one in the bureaucracy can say that publicly. Fighting the War on Terror is now the criterion by which every department has to get more funds, by saying it’s important for the War on Terror. And we saw with Katrina, and we saw with the Iraq war, how the top echelons can cook evidence and twist organizations and bureaucracies so that even professionals who have their hearts and their heads in the right places cannot push the country in the right direction.
  4. And I thought my mom was the only who said this! What about not cooking in non-stick cookware

    Yes, I was forced to throw out all my non-stick cookware (I cook a lot) and had to replace it with stainless steel which food sticks to and is harder to clean as you probably know as well 🙂

  5. my mummy has taken it one step further and now barrages me with the previous night’s happenings on “law and order.” she feels that her daughter living alone in new york is absolutely going to fall prey to every story shown. so my mornings begin with her advising me on how to react if ever faced in those situations.

  6. My mom:

    (During the winter season):Dont sleep with all the doors and windows completely shut or you will die. Dont sleep with the air conditioning on in the car or you will die. Dont go to areas with a lot of those people (insert black) or you will get robbed or die.

  7. everyone who is interested in the “misunderstanding” of statistics should check out heuristics and biases, which suggests that the mind is “hardwired” toward interpreting data in a particular way. in fact, there is research which points to the possibility that quantitative scientists who are good at bayesian probability within their own field get really off base once they are outside of their field, implying that intuitive references that are learned are as important as bayes’ rule in forming a proper mental model. the scariest stuff, for me at least, was how bad medical doctors are at bayesian probability, since it is an implicit part of diagnosis, test and treatment.

  9. here is an aphorism which isn’t strictly true, but get’s to my current frustration with american politics: the left doesn’t like rigor and quantitation in domestic policy and the right doesn’t like rigor and quantitation in foreign policy.

  10. Michael Moore also wrote in Dude, wheres my country? that the chance of falling victim to a terrorist attack in 2001 was less than the frequency of suicides. So in essence, you were a bigger threat to yourself than Al Qaeda. But…, I think its the density that matters, 2000+ people in a span of 1 hour is a lot to inspire fear.

  11. Michael Moore also wrote in Dude, wheres my country? that the chance of falling victim to a terrorist attack in 2001 was less than the frequency of suicides.

    The paper actually comments on a Moore quote, you should check it out.

  12. While living life in massive fear of a terrorist attack and thinking about the world in terms of pre and post 9/11 is counterproductive, dismissing the potential severity of a chemical, biological, or radiological attack is equally misguided. VX gas is a hundred times more potent than sarin, and a dirty bomb attack could totally shut down a major city. How do you know what the potential is for a terrorist act in the next month? How do you know we can do little against terrorism? Have you done the analysis?

  13. How do you know we can do little against terrorism? Have you done the analysis?

    No, I’m too busy blogging. But as this article clearly articulates, people HAVE done the analysis.

  14. A statistician is not qualified to perform an analysis of the possibility of a future terrorist threat. If they could they would also be able to predict the stock market.

  15. However, while obviously deeply tragic for those directly involved, the number of people worldwide who die as a result of international terrorism is generally only a few hundred a year, tiny compared to the numbers who die in most civil wars or from automobile accidents.

    I guess we should also factor in the ‘avoidable factor’. You might be more likely to die in a car accident than in a plane crash, but for most people riding in cars is a necessity while flying in a plane is something which is ‘avoidable’.

  16. A statistician is not qualified to perform an analysis of the possibility of a future terrorist threat. If they could they would also be able to predict the stock market.

    no. we don’t know the atomic unit of economics, all we have to go on are post facto inductive truths (sorry, don’t buy the rational actor that much). an analysis of terrorism is possible because we know the atomic unit here, the human mind, and the interlocking higher order social dynamics.

    to make it short, the stock market is more complex than terrorist threats, there are more parameters (“threat” is not intent but the ability to use WMDs or something like 9/11).

    p.s. if you can’t analyze, perhaps we should consult the astrologers?

  18. A statistician is not qualified to perform an analysis of the possibility of a future terrorist threat.

    Have you ever heard of a probability risk assesment? If not you should become familiar with one since that is how a great many decisions in this world (especially engineering decisions that affect you much more significantly than terrorism) are made. Statistics do predict the future well enough for many of the decisions society makes to be based upon them.

  19. Then again while we’re pondering when the next 9/11, 3/11, 7/11 will occur, there is the possibility of a June 30, 1908 or something like it happening in the future. Time to build that bunker…

  20. Awesome Vikram. Tunguska is my favorite party converation! Statistically speaking another Tunguska will happen in this decade since objects of that size strike the Earth on a 100 year frequency. That might explain why I always leave parties alone though.

    Time to build that bunker…

    The true intentions of SM at last revealed.

  21. Michael Moore also wrote in Dude, wheres my country? that the chance of falling victim to a terrorist attack in 2001 was less than the frequency of suicides. So in essence, you were a bigger threat to yourself than Al Qaeda.

    HAW.

  22. you were a bigger threat to yourself than Al Qaeda

    We’re always a bigger threat to ourselves than anything else, but its about control. Me taking a gun to my temple and letting a 9mm fly through is different that someone else making that choice for me.

    Driving is a good example though. We’ve internalized the risk of getting into an accident not of our doing. Yet, the idea of flying makes people shudder even though it is way safer.

    If people are stuck in the body count mode of terrorism, they’re making the same mistakes as those in the military and offensive operations against insurgents have done so in the past. Such attacks are tactical in nature that serve a terrorist’s strategic purpose. Its about propoganda, power, and dominance. Not the act itself, which is merely a tactic. War on Terrorism is as bad of a phrase as War on Drugs (though I’ll be the first to admit having used it as popular reference). War is fought between living entities, not on tactics (terrorism) or materials (drugs).

  23. Tunguska is my favorite party converation! Statistically speaking another Tunguska will happen in this decade since objects of that size strike the Earth on a 100 year frequency.

    The Nasa NEAR page has a listing with potential Earth impact risk.

    If you check out the Google map of the Tunguska site, you can still see a deforested area (zoom to the 100m level)

  24. Razib,

    Now you admit that economics is difficult to model because the rational actor theory isn’t adequate, but on the other hand you believe we have a complete model of the terrorist mind that will enable us to predict the location and intensity of attacks. Economics is fundamentally the science of human social dynamics.

    I did not say it was not possible to analyse potential threats, but merely that statistical analysis is of limited utility.

    Abhi,

    An engineered system truly is something we know the “atomic unit” of, and hence is susceptible to that kind of analysis.

  26. but on the other hand you believe we have a complete model of the terrorist mind that will enable us to predict the location and intensity of attacks

    huh, we don’t have a complete model, and we can’t predict location and intensity of attacks. i’m not talking a deterministic model here. i’m saying that you used a ‘stock market’ as an analog and predictivity as a litmus test. i think both are off the track in regards to the comparison with terrorism and a cost vs. benefit analysis. apples & oranges.

    in any case, are you a psychic to characterize with such specificity what i believe? back to astrology.

  27. to be precise, the key is to establish the boundary conditions of how terrorists might operate, not exact details of when and where. a big argument about an aggressive war on terror is the potential threat of terrorists, and a probabilistic analysis can give us “reasonable” upward bounds on their destructive capabilities. that’s all.

  28. There is no need for a deterministic model. One can tell what the statistical probability is of being struck by lightning at any given point on earth in a given span of time. Or potentially the rise and fall of regional markets.

    Why do you keep mentioning astrology? Good intelligence networks and the destruction of potential threats is enough.

  29. Economics is fundamentally the science of human social dynamics.

    and just to be precise, economics is an important tool, and the rational actor is an important insight. the problem with many social sciences is that they tend to fixate on one model of human nature (‘the rational actor,’ or ‘class conflict,’ etc.) and simply use that to generate the parameters of a model. history, cognitive psychology, sociology, economics, religious studies, genetics, etc. can all play a role in understanding how different groups of humans behave and the range of their potential behavior. unfortunately, this isn’t physics, but we beating expectation and going beyond introspection should be our goal.

    if one says that something is not analyzable, there is no common ground for discourse, though perhaps someone could put up an interpreative dance which expresses their emotional reaction to potential terrorist threats up on youTube.

  30. to be precise, the key is to establish the boundary conditions of how terrorists might operate, not exact details of when and where. a big argument about an aggressive war on terror is the potential threat of terrorists, and a probabilistic analysis can give us “reasonable” upward bounds on their destructive capabilities. that’s all.

    Really? That seems like recepie to be stunned again by terrorists who decide not to adhere to your reasonable upward bounds.

  31. A statistician is not qualified to perform an analysis of the possibility of a future terrorist threat. If they could they would also be able to predict the stock market.

    and

    Why do you keep mentioning astrology? Good intelligence networks and the destruction of potential threats is enough.

    one the one hand you dismiss statistics, but on the other you clear imply the existence of a model with various weights to the parameters (e.g., what is the risk from a given node in the network?). let me be clear: my own hunch is that you don’t have a problem with statistics per se, but the conclusion of this particular analysis.

  32. That seems like recepie to be stunned again by terrorists who decide not to adhere to your reasonable upward bounds.

    how do you know what my upward bounds are? reading my mind again? (actually, i’d leave that to the experts)

  33. Tunguska was awesome, especially with the spray arrangement of trees. I can watch specials on it over and over again.

    Your post reminds me of my parents.

    Do you know there has been an alarming increase in lightning deaths? Now, are the number (and intensity) of lightning strikes the same as they have always been and, because our ancestors were smarter and simply went inside during a storm, were spared more?

  34. Razib,

    I would of course love to have the ability to read minds (and over the internets!), but sadly can only respond to what you have written. And you’re right, I believe statistics is a powerful and useful tool, when all the parameters are known and measurable, and perhaps in some cases where they are not. My problem is with the thrust of this article, which I believe completely ignores the fact that a cunning and determined enemy can find a way to inflict damage that we probably haven’t thought of or have dismissed. Or something. Obviously the chance of me being killed as I go home today is less here than if I were at home in Iraq. But then turning around and saying we should decrease our investment in preventing terrorism based on that lower risk is what I have a problem with.

  36. To this point in history, biological weapons have killed almost no one.

    and up until 911 nobody had flown planes into skyscrapers. How to prevent terrorists form obtaining WMD’s is the central practical issue of our time, and besides the bush doctrine (from spreading democracy to the 1% rule) there is virtually no comprhensive theory on how to achieve this. even our loyal opposition has been relatively mute, and this from the party that gave us the great unifying theories of the cold war: containment and MADD.

    at least meuller admits that this scenario would throw a wrench in his stats, but lamely asserts that “it would take massive efforts and even more stupendous luck for terrorists regularly to visit substantial destruction upon the United States.”

    But there is no way back to a non nuclear world as humans cannot willingly make themselves ignorant. The cold war solution–where one sides abandons its ideology for the other side’s–is really the only long term solution as Bush understands.

    And we are not fighting terrorism, we are fighting an ideology. even if terror is not used, those citizens who live in muslim societies (including france) risk losing their freedoms through even democratic means, if this ideology is not defeated.

  37. And we are not fighting terrorism, we are fighting an ideology.

    ideology is part of it, but not the totality. the analogy i used in another post on one of my blogs is this: to predict someone’s behavior you can use genetics or environment is parameters alone. but a better model is to incorporate genetics & environment and their interactional parameters. understanding ideology is a necessary precondition, but i do not believe it is sufficient, to modeling this problem.

  38. And we are not fighting terrorism, we are fighting an ideology

    I agree with that. I think the global jihadist Salafists (who are probably a miniscule percentage of Salafists worldwide) are definitely ideology driven. In fact, the bombings in Madrid and London came from Al Qaeda inspired independent salafist cells. I fear that the latter will be the norm in decades to come and its going to be much harder to track them down or to prevent them from launching attacks. Of course the independent cells will probably not have the resources (with current technology) to cause too much damage in terms of number of lives. Most centralized Al Qaeda cells like the ones which operate in South East Asia or Afghanistan are probably more lethal but they will eventually get destroyed by the US.

  40. How to prevent terrorists form obtaining WMD’s is the central practical issue of our time,

    In 1977, a failing Princeton undergraduate student in 5 months designed a ten kiloton (same yield as the Nagasaki bomb) atomic bomb based on publically available information in libraries. Total cost of the bomb : $2000. The Pakistani Embassy tried to get a copy of his term paper. Link

  41. In 1977, a failing Princeton undergraduate student in 5 months designed a ten kiloton (same yield as the Nagasaki bomb) atomic bomb based on publically available information in libraries. Total cost of the bomb : $2000.

    Well Vikram, this pretty much makes Prof. Mueller’s paper irrelevent.

  42. manju, if it was that easy going from design to getting all the materials and implementing the construction of the bomb it seems that we should have long ago been the target of nuclear attack (or someone in the world would have). there isn’t a shortage of bitter and desperate terrorists and rebels. my understanding is that design is one thing, but collecting on the materiel and finding the right engineers to build the thing, and then transporting it, are considerably harder. at least for non-states (and even states have a serious problem, elsewise why would iran not have had a nuclear bomb in 1980?).

  43. Well Vikram, this pretty much makes Prof. Mueller’s paper irrelevent.

    Give me a freakin’ break. That “bomb” wasn’t fueled. Any super smart engineer can make an unfueled nuke. The reason it takes countries decades and billions of dollars to create nuclear weapons in the fuel side of things. I’m sure labeling Mueller’s paper “irrelevant” feels good though.

  44. Any super smart engineer can make an unfueled nuke.

    Sorry, that should have said “nuke design” as Razib pointed out. Some of the actual parts are difficult to machine.

  45. RtA:

    manju, if it was that easy going from design to getting all the materials and implementing the construction of the bomb it seems that we should have long ago been the target of nuclear attack

    ok. maybe the princeton case is not all it’s cracked up to be. but we can only assume it’s going to get easier, not harder…especially w/ states working thru private actors (ie iran/hez) or vice versa.

    bush’s so called 1% doctrine, that the US should treat a 1% chance of a terrorist getting a WMD as a 100% certainty, b/c the consequences are so great if it happens strikes me as a more prudnt policy than meullers, although it is technically a secret policy. how to affect the policy is anothher thing but while meuller addresses the scenario he seems to put his head in a statistical sand.

  46. I’m sure labeling Mueller’s paper “irrelevant” feels good though.

    OK wolf, i feal bad. it’s a well written paper and i think it’s worth debating. but i have objections as i stated.

  47. while meuller addresses the scenario he seems to put his head in a statistical sand.

    That’s kind of ironic since the general thrust of his paper (from my read of it) seems to be that people that have irrational fears or gut feelings (like you are demonstrating in your comments) have their heads up their…

  48. That’s kind of ironic since the general thrust of his paper (from my read of it) seems to be that people that have irrational fears or gut feelings (like you are demonstrating in your comments) have their heads up their…

    His stats are based on past performance, which as every market watcher knows, is no indication of future performance. He cuts short his discussion of the doomsday scenario w/ a casually dismissive “it would take stupendous luck” line, when this scenario is the whole issue; and he even admits it would change the statistical calculation.

    no one, including bush/cheney, is using “irrational fears or gut feelings,” though it must feel good to label counter-arguments as such 😉

  50. I just got off the phone Mom. I couldn’t make this up even if I tried. The topic of conversation that she brought up? West Nile virus. The one topic that I hadn’t researched the stats for.