Wednesday, August 23, 2006

Clearly, I Must Be Missing Something

TigerHawk links to an editorial in the course of complaining about people "who [tell] you that we are overreacting to terrorism because the odds of actually dying in a terrorist attack are somewhere between drowning in a bathtub and croaking from an infected hangnail" on the grounds that, as the editorial puts it,
The probability of, say, a nuclear terrorist attack might be tiny but the consequences, the “payoff” as it were, would be huge. It is expectation, not probability, that should determine policy towards terrorism.
"Expectation", here, means "probability times the payoff".

I don't get what the big deal is supposed to be. Of course you need to consider both the probability and the payoff of the evaluated events, but dying in the blast of a nuclear bomb isn't any worse than dying in your bathtub. The whole point of the bathtub drowning comparisons is that the payoff is just as great - you're dead one way or the other - but the precautionary measures taken in case of terrorism are vastly out of proportion to the probability differential between terrorism deaths and bathtub drownings.

What am I missing?

15 comments:

Tommaso Sciortino said...

I think you are misinterpretting them a little. I think what they are arguing is this:

"Some people say not to worry about terrorism becuase

probability of terrorist act occuring < probability of dying in a bath tub

but really they should note that:

(pb of terrorist act occuring)*(number of people who would be killed) > (pb of dying in bathtub)*(the single individual killed in each case)"

As it so happenes I think that TigerHawk and Finkelstien are just plain wrong here, even if you throw in - as they probably do - the other bad effects terrorism has beyond just killing people. And this isn't some kind of agree to disagree difference of opinion. I think tigerhawk and finklestien don't really know what they're talking about.

Paul said...

OK, that helps me make a little more sense of the editorial. But what's the point of that comparison supposed to be? I would think the relevant comparison is between the overall dangers of terrorism vs. the overall dangers of bathtubs, since we tend to put in place precautionary measures that are aimed at preventing the phenomenon generally, not one specific occurence of it.

So I still feel like I'm missing an argumentative step - their risk assessment seems to me to be relevant only if we are trying to divide resources between preventing a particular terrorist attack and preventing a particular bathtub drowning.

Bill said...

It sounds to me as if Tigerhawk and Finkelstein are simply echoing Cheney's "One Percent Doctrine", which Ron Suskind has explored in his recent book. It didn't make sense when Cheney expounded it, and it doesn't make sense now. Unfortunately, unreasonable arguments seem to draw more support than reasonable ones in our culture.

Tommaso Sciortino said...

Yeah, I think I'll echo what thinker said: You aren't actually missing anything, Paul. They're missing something: a whole logical step which just isn't supported by evidence.

Tommaso Sciortino said...

Don't consider it a wrench. It's more like a circus peanut to be smashed by the iron-like gears of our logical minds.

If a new kind of bathtub came out which caused 3000 deaths in one year I would definitly consider that cause for government action. Just in the same way I consider 9/11 to be cause for government action. The question is, how do we decide how much action? Surely there are negatives to terrorism beyond just the deaths involved - negatives not present in something like bathtub deaths - but we can't just wave our hands and say "synchronized deaths are worse than an equal number of unsynchornized ones".

Bill said...

Aaron, I don't think the question is (or was) whether we (the government) should do something about it. The question is whether what we do increases or decreases risk. If the governments actions decreased the risk of death by terrorism from 1 in a million to 1/8 in a million while increasing the risk of death from poverty or global warming, clearly that is not a net gain, but an overall loss. And in fact, I believe the number of people who've died from acts of terrorism committed in the US since 9/12/01 is zero. However, the number dying from the destruction of the effectiveness of FEMA (which clearly is less effective this decade than it was in the last - before the focus of government changed to all terrorism all the time) is near 2,000 or more. Are we as a nation better off because the Bush Administration is more focused on anti-terrorism than on emergency management? I don't think so. Similarly with their propensity to ignore efforts to offset Global Warming. Changing focus and allocation of resources (time, money, labor, etc.) from those problems more likely to cause calamity to those less likely to do so is clearly unwise and dangerous. On the whole, it makes us less safe (as individuals and as a society).

dasht said...

The statistics concept you need here is independence. Two bathtub accidents are independent. If my neighbor suffers a bathtub accident that doesn't raise the probability that I will share his fate. On the other hand, terrorism breeds terrorism: if my cousin dies in a terrorist attack the odds that I will die similarly gets multiplied.


I made a longer post about this.

Tommaso Sciortino said...

I don't think statistical independence is qutite the word you're looking for. Whatever you call it, you do have half a point: the future also matter. Obviously 3000 deaths due to a contagious virus requires a larger response than 3000 bathtub deaths.

Still. That doesn't make tigerhawks original argument any more valid: there is not a big different between synchronized deaths and unsynchronized ones.

As for your suggested response to terrorism: "smashing systems that reward the families of terrorists; humiliating proponents of terrorism with utter political and economic defeat; spreading peaceful means of achieving political and economic aims such as democracy and trade" I'll say: sure, but no one really disagrees with you. The current political controversies is about how to achieve those goals. Some insane people think invading Iraq and Iran is the way to achieve those goals, leveler heads disagree.

Bill said...

Aaron, recent acts of domestic terrorism prior to 9-11-01 were primarily (if not exclusively) from domestic right-wing groups and individuals (anti-abortion bombers, racists and Timothy McVeigh). As far as I know, no acts of foreign sponsored terrorism took place, much less caused death before or since. The government under Clinton clearly committed time and money to counter terrorist efforts. They successfully stopped potential terrorist acts around the turn of the century. One could make a good argument that had GW Bush been as hands on as Clinton, he would have prodded the bureaucracy to coordinate once he saw the famous August 6, 2001 Presidential Daily Briefing, and 9-11 would have been averted too. There is no evidence I'm aware of that the increased levels of time, spending, and government re-organization of the past four years has made us any safer. And, if the testimony of people like Richard Clarke is to be believed, we are actually less safe as a result.

Paul said...

"Hume's fork" refers to Hume's distinction between analytic and synthetic knowledge; that is, between things we know a priori and things we can know only a posteriori. Hume also said a good bit about inductive inferences about the future (a subset of the judgments from the a posteriori tine of the fork), but his position was that they were ultimately unjustifiable. That particular conclusion is difficult to work with on a day-to-day basis, so we tend not to pay it any mind.

As for the actual inferences in question, I guess I always figured that it was legitimate enough to suggest that the probability of, say, a terrorist attack, might increase in the future, but the questions are: by how much, and on what basis are we making this inference? An attack from Mars seems wildly improbable to me, but a many-fold increase in the incidence of terrorist attacks doesn't seem all that much more probable. (And I don't think that it matters much whether any previous measures taken have held the probability of terrorist attacks down; the issue is just the probability as of today.)

Bill said...

Aaron, since there were no foreign terrorist attacks on US soil prior to 9/11, and there have been none since, what evidence is there that americans are safer now because of administration policies than they were prior to 9/11?

Paul said...

I think Thinker's question is a fair one, but I also think it really doesn't matter, for risk assessment purposes, whether terrorism is being held down by precautionary policies as long as we're talking about whether we need to take additional precautionary measures. Invading Iran, say, or prohibiting liquids on airplanes. It might be relevant if we're talking about scaling back measures that are already in place.

But what does any of this have to do with the wisdom of anxiety about terrorism? What if previous measures were put in place because the President read some precautionary ideas in a fortune cookie? Who cares where the motivation came from?

If the suggestion is supposed to be that the success of previous measures justifies previous anxiety levels, that seems pretty clearly wrong to me.

If people have anxiety about taking baths because of the risk of drowning in bathtubs, and convince Congress to ban bathing, they may very well succeed in reducing the number of bathtub drownings. And we could also say, if we were being charitable, that their anxiety was "well-placed", if for no other reason than that drowning in a bathtub would, in fact, be pretty unfortunate.

A more typical use of "well-placed", here, though, would take into account the fact that it's extraordinarily unlikely that you will die in a bathtub - so it's really not appropriate to be actively anxious about it. Usually, anxiety is not well-placed if it's allocated out of proportion to the relative risk of the event about which one is anxious. The fact that there is some risk doesn't go very far toward justifying anxiety.

Not that much turns on this.

Bill said...

I agree Paul, people's anxieties and fears (and the manipulation therof) are of major concern. In fact, they are the topic of a book by Robert Ornstein and Paul Ehrlich titled NEW WORLD NEW MIND (http://www.amazon.com/gp/product/1883536243/sr=1-1/qid=1156486637/ref=pd_bbs_1/002-0311842-0175223?ie=UTF8&s=books). In it the authors question the psychological mechanisms that humans evolved to enable us to survive in forests, jungles, plains, savanahs, etc., and the limitations they create for us as we attempt to survive in the "civilized" environments we've created for ourselves more recently. Their point, I think, is that we need to develop better ways of testing reality so that we can identify and cope with those things of most real threat to ourselves and our way of life.

Tommaso Sciortino said...

I've only skimmed it. how many of the 858 pages have you read, Aaron?

Be really I don't know what you're responding to. If your just saying that there's a lot of reasonable steps to prevent terrorism that we haven't yet taken then I agree 100%.

This whole thread is kind of weird. The problem isn't the size of the response to 9/11. The problem is the effectivness of that response. Whether we think the response was strategically poor or just executed poorly, we all agree that the response was no good.

Bill said...

Aaron, have you listened to what the co-chairs of the 9/11 Commission have been saying this month about the attack and our response?