Perceptions of Risk

Another article about risk perception, and why we worry about the wrong things:

Newsrooms are full of English majors who acknowledge that they are not good at math, but still rush to make confident pronouncements about a global-warming "crisis" and the coming of bird flu.

Bird flu was called the No. 1 threat to the world. But bird flu has killed no one in America, while regular flu -- the boring kind -- kills tens of thousands. New York City internist Marc Siegel says that after the media hype, his patients didn't want to hear that.

"I say, 'You need a flu shot.' You know the regular flu is killing 36,000 per year. They say, 'Don't talk to me about regular flu. What about bird flu?'"

Here's another example. What do you think is more dangerous, a house with a pool or a house with a gun? When, for "20/20," I asked some kids, all said the house with the gun is more dangerous. I'm sure their parents would agree. Yet a child is 100 times more likely to die in a swimming pool than in a gun accident.

Parents don't know that partly because the media hate guns and gun accidents make bigger headlines. Ask yourself which incident would be more likely to be covered on TV.

Media exposure clouds our judgment about real-life odds. Of course, it doesn't help that viewers are as ignorant about probability as reporters are.

Much of what's written here I've said previously, and it echoes this article from Time Magazine (and also this great op-ed from the Los Angeles Times).

EDITED TO ADD (7/13): A great graphic.

Posted on August 22, 2007 at 1:43 PM • 69 Comments

Comments

tgoAugust 22, 2007 1:54 PM

"Here's another example. What do you think is more dangerous, a house with a pool or a house with a gun? When, for "20/20," I asked some kids, all said the house with the gun is more dangerous. I'm sure their parents would agree. Yet a child is 100 times more likely to die in a swimming pool than in a gun accident."

I'm with you on this one, but there is a statistics problem here. It is not useful to compare the rate of pool accidents to the rate of gun accidents in order to weigh the risk of a pool-equipped house versus that of a gun-equipped house. The data needs to be normed for rate of gun accidents *in gun-equipped houses*, versus the rate of pool accidents *in pool-equipped houses*.

I have no idea what the answer is, but this additional step is needed in order to find out.

aeschylusAugust 22, 2007 1:58 PM

Another thing that doesn't help, in my opinion, is the way many statistical terms have been coopted into regular usage in non-statistical ways.

"It was probably a gun." When in fact, no one has any information about the statistical likelihood that a gun was involved.

"It's normal to do this." No one knows what the norms actually are.

"It's most likely that." Again, no statistical evidence.

... ad nauseam with "most", "usually", etc. All these terms have real numeric meanings, but people use them frequently to mean "I think that...", and in so doing, seem to be trying to borrow a little bit of mathematical credibility where none is warranted. Even the way I just used "frequently" borders on it, though I don't think there's any dispute about that. Once you start looking for this I think you'll notice how rampant it is.

This is perhaps more of a symptom of the same psychological phenomenon that Bruce is referring to than the other way around. (Note how I wrote "perhaps" and not "probably".)

NYCAugust 22, 2007 2:14 PM

Ironically the author of that piece (John Stossel) has a fairly checkered reputation himself. In 2000 the New York Times reported he fabricated evidence in a 20/20 report on organic food. He's also made a number of unsubstantiated claims similar to these ones over the years, but rarely defends his claims or quotes his sources.

conspiracy theoristAugust 22, 2007 2:16 PM

"Bird flu was called the No. 1 threat to the world. But bird flu has killed no one in America, while regular flu -- the boring kind -- kills tens of thousands. New York City internist Marc Siegel says that after the media hype, his patients didn't want to hear that."

I was exceedingly frustrated several years ago when West Nile Virus first became a huge issue in New York City. I kept wondering when the media would stop focusing on the 3 people out of 8 million who had contracted West Nile and start asking why flu vaccines couldn't be made more accessible. To this day, I am convinced that Giuliani whipped the city into a pesticide-demanding panic for some bizarre political reason of his own.

Vincent GableAugust 22, 2007 2:17 PM

@ tgo, you are absolutly correct. As near as I can tell, the statistics comes from a book "Freakonomics", and found that "one child under 10 drowns annually for every 11,000 pools. By comparison, one child under 10 each year is killed by a gun for every 1 million guns"

Source:
http://www.alphecca.com/mt_alphecca_archives/...

A few problems: in my experience, most households that own one gun own several, while nobody has more then one swiming pool in their house. Also, this is only children under 10, and my understanding is that gun-deaths are more likely to occur with older children/teenagers. All of these factors lead to over-counting.

From the blog:
In Arizona:
— Accidental deaths in Arizona for children, 2000-2003
— Drowning: 140
— Gunshot wound: 15
Source: Arizona Child Fatality Review Program

Which is a factor of 10 under the 100x claimed.

Unfortunately, finding the number of households with guns is somewhat tricky, since AFAIK information (background checks, receipts, etc) is per gun owner, not per household.

I personally think the conclusion (pools more dangerous then guns) is valid, but I agree that the 100x figure is highly suspect.

chipuniAugust 22, 2007 2:17 PM

This material was also covered by the book Freakonomics. It's not new in the current public sphere.

I'd be curious to see how deep math illiteracy goes. Does anyone out there have good studies (obviously not media reports of studies!) that talk about the current state of math illiteracy (John Allen Paulos calls it "innumeracy") in the US?

Erik NAugust 22, 2007 2:21 PM

"What do you think is more dangerous, a house with a pool or a house with a gun? When, for "20/20," I asked some kids, all said the house with the gun is more dangerous. I'm sure their parents would agree. Yet a child is 100 times more likely to die in a swimming pool than in a gun accident."

Yeah, and even when numbers are correct they don't know how to interpret them. Obviously, from this quote the author wants to say: We should not regulate guns because swimming pools are more dangerous, and obviously we can't prohibit swimming pools.

But, you're not supposed to mix guns and kids, people understand that and avoid to do so. Hence less accidents happens. But, swimming pools are largely made for the kids, accidents are bound to happen more often.

Comparing non-related risks can be just as dangerous as misinterpreting numbers.

ThoriaAugust 22, 2007 2:22 PM

@aeschylus,

*Perhaps* it's because all those words had common meanings before the science of statistics was developed, and the o-opting happened in the other direction. "Probable", for instance, comes from Old French.

Patrick HenryAugust 22, 2007 2:29 PM

@Erik N

At least in NJ, swimming pools are regulated. Fences are required around either the pool or the property.


Purely speculation, but I always figured a house w/a swimming pool is more dangerous than a house with a gun or guns because swimming pools are used, often and on purpose, by children and because minor accidents with pools are probably extremely frequent while most accidents with guns are extremely rare but serious.

aeschylusAugust 22, 2007 2:31 PM

Thoria> "Probable", for instance, comes from Old French.

Naturally the words weren't made up out of whole cloth, and they have a distinct etymological meaning that existed before the mathematical science of statistics.

Nevertheless, pull someone aside and ask him or her what "probable", "most", "usual" mean, specifically, and I will bet you real money that you will *usually* get a statistical answer, e.g. "more than 50% likely", "more than half", "more often than not", etc.

DMAugust 22, 2007 2:35 PM

The figure of pools being 100 times more dangerous than guns is probably based on a study by Gary Kleck, cited in http://www.guncite.com/gun_control_gcgvacci.html, in which the correction for the fact that guns are 8 times more prevalent than pools was taken into account.

But the comparison is still faulty. It restricts itself to accidental deaths. Most gun-related deaths are not accidental. If you want to know whether it's safer to visit a house with a gun or a pool, you should compare the rates of deaths of all kinds.

In the US, there are about 2000 pool drownings per year, and about 30000 firearm-related deaths.

But that's still not right --- many of those drownings and shootings don't occur in homes. But then some of the firearms used elsewhere were stolen from homes...

GretchenAugust 22, 2007 2:38 PM

Wow, made the same error he's arguing against.

Pandemic flu is a low probability, high consequence event. Like a hurricane flooding New Orleans and causing the levies to break. We knew that it could happen, and even that it would happen, eventually.

To say that these things are not "real dangers" but "things that hardly matter" is very misleading.

Pandemic flu is a real danger. And when it arrives (which it will, with 100% certainty) it will matter a great deal. To say that we should ignore this risk in favor of seasonal flu is as wrong as saying it the other way around.

It is exactly this sort of flippant reporting that does the harm Stossel is compaining about.

JohnSAugust 22, 2007 2:38 PM

re: perceptions of risk, see also
Sunstein, C. R. (2005). Laws of fear: Beyond the precautionary principle. Cambridge,UK and New York, NY. Cambridge University Press. and generally Kahan, D. M. & Braman, D. at the Cultural Cognition Project.

bobAugust 22, 2007 2:41 PM

@conspiracy theorist: He's merely taking a page (so to speak) out of "Silent Spring"; the novel which got DDT banned in most of the western world subsequently contributing to [literally] millions of human deaths in the developing world through mosquito-borne diseases, in order to save birds - without any actual science proving a relationship between birds and DDT.

Joe PattersonAugust 22, 2007 2:53 PM

People don't seem to realize that "news" is focussed on the extraordinary. No one reports on the normal, they report things that are unusual. Thus, if you see something in the news, it should be *less* of a concern for you. If it were really dangerous to most people, then it wouldn't be newsworthy. It's an emergent effect of widespread communication. I can't know all things about all 6 billion people on the earth, nor can anyone else. So in our communications with one another, we focus on the extraordinary (mark that word carefully: Not Ordinary). Thus we build up a worldview where the people close enough to us that we *can* know a great deal about their lives seem fairly true-to-life, but the rest of the world seems filled with freaks and wild dangers. It's like attempting to do a statistical analysis by throwing out all data except for the outliers.

Carlo GrazianiAugust 22, 2007 2:56 PM

Thoria says: "*Perhaps* it's because all those words had common meanings before the science of statistics was developed, and the o-opting happened in the other direction. "Probable", for instance, comes from Old French."

Yes, but nowadays the word "Probable" is used exclusively in a manner that is totally different from its usage prior to the rise of probability theory.

"Probable" is cognate to "approvable". Its original usage was to characterize opinions based on the authority of those who held them. That is, an opinion was held "probable" if it was expressed by someone in authority, or by someone held in high esteem. It had no real relation to the notion of intrinsic plausibility. Innumerates who use the term today are not using it in the medieval sense. They are using it in the modern sense, just without any technical justification.

Erik NAugust 22, 2007 2:57 PM

According to statistics, living organisms run a significant risk of dying. Life is much more dangerous than guns, the only reasonable measure is to prohibit life itself and kill everyone.

yoshiAugust 22, 2007 3:03 PM

@gretchen

warming - rant starting....

Here is the problem with your argument. According to the media and "authorities" every year I am going to die of something with 100% certainty. I was "taught" that nuclear war was certain in the 80s. I was "taught" that oil would run out by year 2000 and we be dead because of massive energy wars. I was "taught" that food will disappear and we all be dead of starvation. Now I am being "taught" that climate change and/or pandemic flu will wipe us all out (or a significant majority). Predictably I look at these so call 100% certain predictions with a heavy degree of skepticism.

Its your kind of flippant reporting that we should be worried about.

rant ended...

I take appropriate measures to handle risk in myself and my families lives and part of that is not believing every sensationalist story out there.

qkslvrwolfAugust 22, 2007 3:06 PM

I was going to ask, isn't there some need to identify *potential* for harm as well as actual rates of harm? Risk assessment, IIRC, involves not only finding out how damage something will do IF it happens, HOW LIKELY is it to happen, and weighing that total against the cost of preventative measures, or some such.

I also think that the guy's virtual finger quotes around global-climate change doesn't exactly speak highly of his ability to read scientific reports, since any number of bodies that I consider fairly authoritative (like the UN's Intergovernmental Panel on Climate Change IPCC) have said that there is a very real risk there. Unless the guy is among those claiming that scientists like to make shit up, as a community, just to ensure funding, I don't think that the quotes were necessary.

AndrewAugust 22, 2007 3:07 PM

Stossel needs to brush up on his math. As another commented pointed on, it's not the probability, it's the expected value. in this case, the expected value of harm. A pandemic is low probability for any particular year, but is high-impact. Multiply probability by number of deaths (or whatever metric you wish).

And the scare quotes around the global warming "crisis"? Is John Stossel, reporter, saying that he can judge the risk greater than climatoligists with some expert in the matter?

There is also some unquantifiables. It may be logical to be scared of airplane crashes rather than heart attacks. Maybe dying in an airplane crash is 1000 times worse than dying of a heart attack....

Other than those oversights, his other points are valid.

qkslvrwolfAugust 22, 2007 3:09 PM

I really gotta use the preview button.

I'm basically asking what the "formula" for risk assessment is, because I recall it involving an assessment of likelihood of an event combined with the damage of that event to create a justification for the costs of preventing or mitigating that event.

Does that make any more sense?

I stand by the climate change bit, though.

platypusAugust 22, 2007 3:10 PM

@yoshi:
"According to the media and 'authorities' every year I am going to die of something with 100% certainty."

Citations please? What media figures and "authorities" have said that this year? Last year?

aeschylusAugust 22, 2007 3:11 PM

Carlo Graziani> Innumerates who use the term today are not using it in the medieval sense.

I'll award Thoria a point, in retrospect, in that "coopted" was not the most precise word I might have chosen.

Excellent use of "innumerates", by the way--such a fun word. ;^)

AnonymousAugust 22, 2007 3:21 PM

John Stossel? Might as well quote John Lott.

Stossel has a tendency to manipulate the numbers to fit his theories. He doesn't deserve to be linked to from here.

BunnyAugust 22, 2007 3:33 PM

"because the media hate guns"

...yes. Sorry, but while the article still may have some good points, it lost me there.

Rionn Fears MalechemAugust 22, 2007 3:34 PM

Andrew,
Excellent call bringing out Stossel's scare quotes around the "global-warming 'crisis'" and the false implication that it's not the number one threat to civilization.
Ion

bobAugust 22, 2007 3:37 PM

@DM: Further you should mention that while very few of the pool deaths were "intentional", the majority of deaths by firearm were - a policeman shooting a felon, for example; not exactly something that one would want to reduce per se.

kpAugust 22, 2007 3:37 PM

Boy, Stossel's agenda is showing a bit, isn't it. Nothing he says is outright wrong, except maybe the probability * outcome noted above, but he consistently cherry-picks examples against traditional liberal causes: gun control, global warming (which didn't fit with any part of his thesis anyway) and of course The Media.

Actually Bruce, I think his railing against the media clouds the point you are trying to make: the media doesn't make us think this way, it's the way our brains are wired: focus on the spectacular. Gun deaths spur a more visceral reaction in people, so gun deaths sell more papers, so they get reported more.

Craigh HughesAugust 22, 2007 4:01 PM

"Bird flu was called the No. 1 threat to the world. But bird flu has killed no one in America, while regular flu -- the boring kind -- kills tens of thousands."

This is of course bullshit. The "30,000 deaths per year from influenza" number which is always bandied about is a mean number, including the 1917-1918 pandemic. The median number since about 1970 is more like 1,000, about 99% of whom are age 65 or older, and the vast majority of those were already infirm in one way or another.

If 30,000 Americans truly died "every year", you'd probably know of someone who died of flu -- an uncle; a co-worker's half-brother... Who does?

I've seen the "30,000 deaths on average per year" mutated by reporters who wouldn't know what "arithmetic mean" meant if you explained it to them, as variously such things as "30,000 deaths every year" or "You know the regular flu is killing 36,000 per year". It's complete bullshit. Spanish Influenza killed a whole crapload of people 90 years ago, a few more about 40 years ago, and "regular flu" kills almost nobody, ever.

Pandemics are something worth planning for, though probably not for any specific pandemic, but rather just generic massive medical emergency readiness plan of some kind. It might not be avian flu -- it might be mutant pig-hybrid HIV, or typhoidal ebola mumps. But generally having some plan in case of massive lethal viral outbreak would probably not be a bad idea. Giving a crap about something which only kills 1,000 people per year in the US, most of whom probably would have died of something else if they hadn't caught the flu, is kind of a waste of everyone's time and newspaper column inches.

Kevin WayAugust 22, 2007 4:03 PM

I'm deeply disappointed that you would link to a John Stossel article. He's an idealogue, not truth-seeker.

He poo-poos global warming by claiming that the alarm is coming from English majors who are no good at math, which simply isn't true. The English majors are reporting on what is being said by thousands of scientists, all of whom are very good at math.

He makes some valid points, but honestly... John Stossel? You might as well link to Penn & Teller's Bullshit.

AnonymousAugust 22, 2007 4:08 PM

Of course, the influenza vaccine itself can cause adverse reactions, including death.

So what do we focus on?

Heart disease, cancer, stroke: Let's eradicate them or at least delay onset for another thirty or forty years of age.

Driving under the influence: You should get shot in the knee after eating raw "bird flu" infected chicken and get dropped into a pool that doesn't have a drain cover.

Carlo GrazianiAugust 22, 2007 4:15 PM

Of course, widespread inability to calculate realistic risk is one of those "asymmetric information" situations that can create market opportunities for the better-informed.

In my more PT Barnum-influenced moments, I wonder how wealthy one could become selling alien abduction insurance.

GretchenAugust 22, 2007 4:52 PM

@yoshi

Here's the problem with YOUR argument - overestimating risk can be just as deadly as underestimating it. Because nuclear war didn't happen, should we conclude it is just a "sensationalist story?" Of course not, there is real threat to nuclear proliferation and all that goes with it. Concluding that because there is a risk of nuclear war you should worry about that exclusively and not get your flu shot is foolish. Conversely, to get your flu shot and determine that nuclear war is "not a real danger" is also very foolish.

I hear the same about Y2K and how it didn't happen and therefore was all some sort hyped up media lie. True, much of the coverage about consequences was speculative and way over blown, but the underlying substance was true. In the end, Y2K was of little consequence because many people put in several years of work ensuring that disruptions wouldn't happen. I was one of them. Then everyone concluded that because it didn't happen it had never been a threat. Gheesh. You're welcome.

I live in Minneapolis. We don't have bridges collapse everyday, but when it happens, we pay attention. We then ask ourself why we didn't pay attention before. Maybe if we had taken action before the bridge collapsed, we could have prevented it. But, perhaps we said it was "not a real danger" or one of those "things that hardly matter." Maybe we thought that because seasonal flu kills more Minnesotans that collapsing bridges, we should only be concerned about getting flu shots to people and not worry about the bridges. We should worry about both, and take appropriate action on both.

Stossel's point (I think) is that the media is a bad place to turn to develop a true understanding about risk, because the reporters aren't good at risk communication. Or something like that. He then goes on to demonstrate his point by telling us to ignore events that do have substantial risk (although low probability) associated with them. The Chicken Littles who taught you obviously went to the Stossel school of risk communication.

n0_j0August 22, 2007 5:10 PM

@yoshi:
"Now I am being "taught" that climate change and/or pandemic flu will wipe us all out (or a significant majority)."

No, you aren't being taught that. What people are saying is that these are some problems that we see a high probability of happening sometime soon (within your lifetime), and if we don't react appropriately, we could regret it. If we do act appropriately, the effects could be minor (and then you'll go ahead and complain that there are two more things that you were "told" would wipe us out and didn't).

Do you think that Y2K wasn't a big problem because it was all a big trumped-up scare tactic, or do you think it wasn't a big problem because lots of people put lots of time into fixing the problems before they caused trouble?

Robert BAugust 22, 2007 5:25 PM

I am tired of being manipulated and lied to!!!!!!

For example, take the case of second-hand smoke in the workplace/bars/airports claimed by the cancer society, EPA, and every other idiot who has rallied behind this sheer stupidity. This is also a case of "Tyrrany of the majority"

Keeping a major study under wraps

Conspicuously missing in the position statement is any reference to a 1998 World Health Organization (WHO) study, one of the largest ever done on environmental tobacco smoke and lung cancer risk. Overall findings from this study, published in the Journal of the National Cancer Institute, show no statistically significant increased lung cancer risk for nonsmokers exposed to environmental tobacco smoke in the home, the workplace, vehicles, or public places such as restaurants.

As acknowledged in the WHO’s own press release (3-9-98), a small increased risk was found for nonsmokers living or working with smokers, but “neither increased risk was statistically significant.��?

Furthermore, in contrast to the current anti-smoking hysteria to ban smoking in restaurants, the WHO study states that “public indoor settings [including restaurants] did not represent an important source of ETS exposure.��?

Giving credence to a discredited report

Referencing the U.S. Environmental Protection Agency’s 1993 report, Ducatman and McLellan say that the EPA “has estimated that 3,000 excess lung cancer deaths occur yearly among non-smokers due to environmental tobacco smoke.��? What Ducatman and McLellan don’t disclose is that the EPA’s findings were ruled null and void in 1998 by U.S. District Judge William Osteen who found that the Agency manipulated the data in order to arrive at a desired and predetermined conclusion.

As reported in the Washington Post, July 19, 1998, “A federal judge has ruled that the Environmental Protection Agency wrongly declared secondhand tobacco smoke a dangerous carcinogen in a landmark 1993 report, a decision that could imperil hundreds of local and regional ordinances banning indoor smoking.��?

http://www.junkscience.com/feb01/perske.htm

All I want is to be able to have a cigarrette in the airport, or a bar. I'm not asking for much, I just want to be able to have a cig outside on the sidewalk, be able to smoke in one or two bars in an airport.

Think we're living in a police state now??? Try being a smoker!
I am soooooo tired of the American people being treated as sheep, and to see us all act like sheep! It's unforgivable.

I joined the Army when I got out of high school, not just for the college money, but because I believed in protecting our freedom. Friends of mine joined for the same reason, friends that have died in service to our country. To see these freedoms (civil liberties of all kinds) yanked away by these fools in Washington and local government makes me very sad. It is disrespectful to our fallen soldiers and a disgrace to everything this country was built on. I am insulted.

AnonymousAugust 22, 2007 5:35 PM

All I can say is I'm relieved no terrorist have tried to smuggle a swimming pool onto an airplane. The carnage. The devastation. Oh... the humanity!

TheOtherAndrewAugust 22, 2007 5:46 PM

Comparing apples and oranges. Children drown in swimming pools largely by accident. Children die of gunshot wounds due to murder, suicide, and negligence (would include "playing" with a firearm by a child too young to understand). Fences and gun locks may be required, but can be climbed and unlocked. Swimming pool covers and safes provide additional protection, but not foolproof. I'd want to see some real statistical analysis having to do with children who know how to swim; children who have taken "Eddie Eagle" firearms safety training; and last but not least, a good urban-rural comparison to weed out hunting accidents from drive-by shootings.

When I took a class taught by a risk management specialist about ten years ago, we learned that risk perceptions are strongly influenced by the illusion of control. People feel safer driving (where the risk of injury or death is high) than flying (where the risk is much lower) because they feel out of control.

Gun owners would feel that they have considerable control over the risk of their children dying in a firearms incident in their home, and some control elsewhere (as they have taught their children to be safe around firearms.) Some parents teach their children to swim, and thus would feel safer about their children swimming ... perhaps too much so, if the children are permitted to swim unattended. People afraid of guns would feel out of control of the risk of their child encountering a firearm in someone else's home, and take steps accordingly (such as not allowing their children to play at the house of a person who owns a gun.)

I think the article is full of poor science, and any connection with truth is a happy but unlikely accident.

Fraidy CatAugust 22, 2007 5:49 PM

Pandemic Flu is a real risk. It might or might not be H5N1 that is the next pandemic flu, but there *will* be another pandemic flu. It might be milder than 1918, or it could be worse. Preparing for a pandemic flu isn't something that can be done overnight, so it makes sense to make long term preparations for an event that is low probability (for a given year) with a large risk. This is not unlike preparing for earthquakes or hurricanes.

Flu Wiki (http://fluwikie.com/) is an excellent resource.

That said, I agree that an annual flu shot and the other "boring" advice from your doctor is worth following.

Pat CahalanAugust 22, 2007 6:04 PM

@ DM

> But the comparison is still faulty. It restricts itself to accidental deaths.
> Most gun-related deaths are not accidental. If you want to know
> whether it's safer to visit a house with a gun or a pool, you should
> compare the rates of deaths of all kinds.

Ah, no. You're sort of headed in the right direction, but now you're into a related-rates question - you're not just looking at statistics, but demographics.

You're correct, from a safety standpoint you may want to take into account willful shootings, but this depends upon what question you're actually asking.

If you're asking "what is *more likely to cause death to a homeowner*, having a gun in the house vs. having a pool on the premises?", you need to count (a) accidental gun deaths in the home (you can't count, for example, accidental gun deaths at a gun range) plus (b) intentional deaths in the home, where the homeowner was the *target* of the shooting (if I'm a homeowner and I'm murdering someone intentionally, the gun ownership isn't a risk, per se), and compare (a+b) to the total number of homes with a gun. Similarly, you'd need to count (a') accidental drownings in the home and (b') intentional drownings in the home and compare (a'+b') to the total number of homes with a pool. Finally, you need to compare them *independently* to the number of people who die from *all causes* in the home per household.

Then you'd also need to estimate how changing one of those rates would have an effect on the other. If your results lead you to ban gun ownership, for example, and there is a sudden uptick in the number of people who are forcibly drowned in backyard pools or are stabbed by their spouses, you haven't improved your safety any, you've just moved the group of "I want to kill people" to a different methodology.

So, you'd need to model "removing guns from society", and see what the effect was on the "deaths in a household" rate. If the effect is negligible, you can't count intentional shootings in your risk analysis -> taking the gun out of the household just led someone to beat someone else to death with a claw hammer, it didn't actually reduce the risk of *intentional death* at all.

I suspect that adding in intentional shootings may make a difference, as it's generally shown that the likelihood of a crime of passion is linked to the availability of a weapon, but once you correct for the murders that will take place regardless of the weapon that is used, I suspect this will have a marginal overall effect in comparing the rates.

AnonymousAugust 22, 2007 6:16 PM

@DM
"In the US, there are about 2000 pool drownings per year, and about 30000 firearm-related deaths."

That's for the total population. The article specifically says "children under 10". What are the numbers for them?

Pat CahalanAugust 22, 2007 6:39 PM

@ Robert B

I'm not sure where your junkscience article gets information regarding the 1998 WHO study, but according to the WHO itself, the 1998 study results are in direct contradiction to what your junkscience article says they are:

http://www.who.int/inf-pr-1998/en/pr98-29.html

See also:

http://www.who.int/tobacco/research/...
http://monographs.iarc.fr/ENG/Monographs/vol83/...

I'll not comment on the federal judge ruling, according to the EPA site they stand by their original report (http://cfpub2.epa.gov/ncea/cfm/recordisplay.cfm?deid=2835).

aetiusAugust 22, 2007 7:36 PM

The most frustrating part of this whole debate is that there is so much presumption that we have any actual control over these numbers at all. Lets say that 30,000 Americans die each year from guns (all causes). If there are 300 million Americans, that's ... .01% of Americans. What makes us think that we can alter this statistic by one order of magnitude either way? Can you imagine the amount of effort it would take to do that?

It's similar to the argument about motorcycle riders and helmets. The difference between the two states (riders with helmets in accidents, and riders without helmets in accidents) is so completely insignificant that is is literally impossible to measure, let alone determine causes, effects, and possible actions. It defies reason that these things are actually argued, let alone legislated. We are having a self-induced allergic reaction to being safe.

MMHAugust 22, 2007 9:04 PM

The gov't and media to great lengths to teach us that we must tolerate no risk, it's just too dangerous. Therefore, we must ban both guns and pools, and we could save all those innocent lives lost every year. [think of the children, people...]

Stefan WagnerAugust 22, 2007 10:34 PM

@ Craigh Hughes:
"It might not be avian flu -- it might be mutant pig-hybrid HIV, or typhoidal ebola mumps."

made my lol.
Still rolling. :)

Thanks for that.

Stu SavoryAugust 22, 2007 11:56 PM

There are more gun deaths in countries whose head of state is titled "President" than in countries where this not the case. Therefore such countries should be banned.

AnonymousAugust 23, 2007 3:30 AM

@Anonymous

The brilliance of the whole article is that it asks the question "What do you think is more dangerous, a house with a pool or a house with a gun?" and leaves us debating the risk to under 10s which are probably exactly the exception. Even better; the question its self is a stand in for the real issue in our heads (are swimming pools more dangerous than guns) to which the answer is approximately 2000:30000 or guns are 15 times more dangerous. Best of all, even the real question is just a distraction from the real issues: a) are the benefits of swimming pools (reduced heart disease; coolness in hot weather) worth the cost (2000 deaths per year)? b) Are the benefits of guns (xxxx many people defending themselves from attacks; yyyy many people not being attacked in the first case etc.) worth the costs (30k deaths; wide availability of tools of intimidation etc.)? c) are there any ways we could change those trade offs at reasonable cost (e.g. swimming pool fences and alarms or weapons that only fire when held properly by their owner and otherwise trigger alarms)?

To be honest; I think the English majors are pretending ignorance in order to manipulate us :-)

TSAugust 23, 2007 9:25 AM

What kids did the author ask? City or suburban? It makes a huge difference. Urban pools have lifeguards on duty, so there will be the perception of safety, and I would guess that they are in fact safer than pools in a suburban home.

Urban kids also have to face higher rates of crime. There are more crimes committed with weapons. So yes, the perception is going to be that a gun is more dangerous.

Kirk UnitAugust 23, 2007 10:11 AM

"Parents don't know that partly because the media hate guns and gun accidents make bigger headlines."

What a preposterous conclusion to jump to. An article claiming to debunk unsupported assertions shouldn't be making such unsupported assertions.

DMAugust 23, 2007 10:19 AM

@Pat Calahan:

> If you're asking "what is *more likely to cause death to a homeowner*,

No, I was trying to come up with the same question the kids were asked. The article says "What do you think is more dangerous, a house with a pool or a house with a gun? When, for '20/20,' I asked some kids, all said the house with the gun is more dangerous."

That question is ambiguous, because it doesn't say "dangerous to whom". So a reasonable interpretation is "dangerous for a kid to visit". I don't think "dangerous to the homeowner" is reasonable, because the question was addressed to kids. Another reasonable interpretation might be "dangerous for a kid to live in".

But the real point of my post was that the ratio in the article based on accidental deaths (guns:pools = 1:100) and the ratio based on total deaths (guns:pools = 15:1) are both invalid comparisons. I agree with you that this is a hard question.

And to Bob, who said "a policeman shooting a felon, for example; not exactly something that one would want to reduce per se." -- I can't see why you wouldn't want to reduce that kind of death, but I agree it shouldn't be counted in the risk of visiting a house with a gun (unless a felon lives there, and the policeman is a bad shot. Not exactly a negligible risk.)

Robert BAugust 23, 2007 11:51 AM

On the Second Hand smoke issue:
It really has been made out to be something it is not. I don't understand why the government of California had to manufacture the science attacking it, but it has, The WHO went along with it, cause, well it seemed legit...

I'd like to see a study comparing second hand tobacco smoke exposure and traffic exaust exposure.

Here's a British Medical Journal Article finding nothing but a negligible impact of environmental tobacco smoke:
http://www.data-yard.net/43/1057.pdf

and another:
http://www.bmj.com/cgi/content/full/320/7232/417

And a German study from Johns Hopkins/ american journal of epidemiology:
http://www.data-yard.net/39/cabin.pdf

and another, yawn...
http://jnci.oxfordjournals.org/cgi/reprint/90/19/...

These are reputable well run medical studies. Do you feel lied to? You should.

antibozoAugust 23, 2007 2:57 PM

Robert B> To see these freedoms (civil liberties of all kinds) yanked away by these fools in Washington and local government makes me very sad.

Just to clarify: are you saying that you believe your inherent civil liberties include the right to administer an airborne stimulant (nicotine) to passersby?

JaredAugust 23, 2007 2:59 PM

> 30,000 Americans die each year from guns ... If there are 300 million Americans, that's ... .01% of Americans

Uh, only 0.83% of Americans die each year, period. So to compare deaths from guns to total population is fairly meaningless. You compare it to total deaths from all causes. In which case it accounts for roughly one in 80 deaths, on par with fatal automobile accidents. Which is rather high compared to, say, terrorism which is at one in 30,000 deaths or so.

Robert BAugust 23, 2007 5:01 PM

@antibozo,

The research has clearly stated that environmental smoking does not constitute "administering a relevant level nicotine" to anyone but the person smoking it. (read the scientific journals)

I challenge you to read the studies. Is passive smoke a danger, I think we can say no; Is it an annoyance to people who are non smokers? Sure, sometimes.

There are plenty of annoying things we put up with in a free society, loud stereos, annoying people. Some people are courteous, some are not. That's life, I am not advocating that people be able to smoke everywhere, but I think there's a middle ground especially if you look at smoke for what it really is: an annoyance to some, and perfectly acceptable to others. I think business owners can make up their own mind what sort of market they'd like to attract.


My point is (along the lines of this blog) is that, among the ranks of avian flu, and other over-blown, and basically dumb "dangers", second hand smoke stands proudly, and I am affected by it.

antibozoAugust 23, 2007 5:42 PM

Disclaimer: don't mean this to sound personal, Robert, but I feel fairly strongly about the general degradation in quality of life wrought by smokers on the rest of us. So tone it down in your head when you read it, k?

Robert B> (read the scientific journals)

Which ones? The ones you've been cherry-picking, or all of them?

Robert B> Is passive smoke a danger, I think we can say no; Is it an annoyance to people who are non smokers? Sure, sometimes.

As someone who is not inured to the effect of tobacco smoke by years of addiction, I can tell you that minor exposure causes a definite physiological effect in those of us who are sensitive. Whether you regard this as a "relevant level of nicotine" or not, it's there, and it's far more than an annoyance to be randomly rendered nauseous and lightheaded by people with little consideration for the well-being of themselves and others.

Robert B> I think business owners can make up their own mind what sort of market they'd like to attract.

As has been amply demonstrated in places such as Ireland, unless someone says "all of you, knock it off," they aren't, since their perception is that their competitors will have a market advantage. It would be interesting to know how many establishments would go back to allowing smoking if these bans were rescinded.

AnonymousAugust 24, 2007 1:17 AM

@Erik:
"swimming pools are largely made for the kids"

Rubbish. I have a swimming pool precisely so that my wife and I *don't* have to share a pool with kids.

antibozoAugust 24, 2007 3:52 AM

Robert B, here are a few articles for your review. I would say I was cherry-picking, but it's hard to call it cherry-picking when you don't run across a single example that disputes your position.

http://jama.ama-assn.org/cgi/content/abstract/...
"Respiratory Symptoms, Pulmonary Function, and Markers of Inflammation Among Bar Workers Before and After a Legislative Ban on Smoking in Public Places"
2006/10/11, Journal of the American Medical Association
"Context: Scotland prohibited smoking in confined public places on March 26, 2006."
"Conclusions: Smoke-free legislation was associated with significant early improvements in symptoms, spirometry measurements, and systemic inflammation of bar workers. Asthmatic bar workers also had reduced airway inflammation and improved quality of life."

http://jnci.oxfordjournals.org/cgi/content/full/...
"Metabolites of a Tobacco-Specific Lung Carcinogen in Nonsmoking Women Exposed to Environmental Tobacco Smoke"
2001/03/07, Journal of the National Cancer Institute
"Background: Environmental tobacco smoke (ETS) is associated with lung cancer in nonsmokers. Most epidemiologic studies find a higher risk for lung cancer in nonsmoking women married to smokers than in those married to nonsmokers. We measured metabolites of a tobacco-specific lung carcinogen in urine from healthy, nonsmoking women exposed to ETS."
"Results: Urinary levels of nicotine, cotinine, NNAL, and NNAL-Gluc were statistically significantly higher in exposed women than in unexposed women."
"Conclusions: Nonsmoking women exposed to ETS take up and metabolize the tobacco-specific lung carcinogen NNK, which could increase their risk of lung cancer. Within couples, the NNAL plus NNAL-Gluc level in exposed women compared with that of their smoking partners averaged 5.6%."

http://jama.ama-assn.org/cgi/content/abstract/...
"Association of Pediatric Dental Caries With Passive Smoking"
2003/03/12, Journal of the American Medical Association
"Context: Dental decay is the most common chronic disease of children and it disproportionately affects those living in poverty, but the reasons for this are not clear. Passive smoking may be a modifiable risk factor for dental caries."
"Conclusions: There is an association between environmental tobacco smoke and risk of caries among children. Reduction of passive smoking is important not only for the prevention of many medical problems, but also for the promotion of children's dental health."

http://jama.ama-assn.org/cgi/content/full/280/22/...
"Bartenders' Respiratory Health After Establishment of Smoke-Free Bars and Taverns"
1998/12/09, Journal of the American Medical Association
"Context: The association between environmental tobacco smoke (ETS) exposure and respiratory symptoms has not been well established in adults."
"Conclusion: Establishment of smoke-free bars and taverns was associated with a rapid improvement of respiratory health."

http://cebp.aacrjournals.org/cgi/content/full/14/...
"Urinary Metabolites of a Tobacco-Specific Lung Carcinogen in Nonsmoking Hospitality Workers"
2005/05, Cancer Epidemiology Biomarkers & Prevention (American Association for Cancer Research)
"Exposure of nonsmokers to environmental tobacco smoke results in increased risk for cancer and other diseases. In spite of this finding, some restaurants and bars continue to permit smoking. This study examined the uptake of nicotine and 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK), a potent lung carcinogen, in nonsmokers who work in restaurants and bars that permitted smoking... The results showed significant increases in urinary levels of total NNAL, total nicotine, and total cotinine on working days compared with nonworking days. The results of this study show that smoke exposure in bars and restaurants may have important health effects on nonsmoking employees, elicited by the increase in carcinogen levels."

http://www.ehponline.org/members/2004/7210/...
"Exposure to Environmental Tobacco Smoke and Cognitive Abilities among U.S. Children and Adolescents"
2005/01, Environmental Health Perspectives
"We used the Third National Health and Nutrition Examination Survey (NHANES III) , conducted from 1988 to 1994, to investigate the relationship between environmental tobacco smoke (ETS) exposure and cognitive abilities among U.S. children and adolescents 6-16 years of age. Serum cotinine was used as a biomarker of ETS exposure."
"The findings of this study confirm previous research indicating an inverse relationship with ETS exposure and cognitive outcomes. We also provide new information indicating that ETS is neurotoxic at extremely low levels. Exposure to ETS in U.S. children therefore has substantial public health impact beyond asthma, otitis media, and other widely recognized adverse consequences."

NMAugust 24, 2007 6:12 PM

Re: Bird Flu

I'm not sure I understand the point about bird flu in the original article. The risk of catching bird flu as it is is extremely low for any individual, at the moment. What's at stake is the /pandemic/ that would occur when a strain of bird flu mutates, crosses the species barrier and proliferates.

Until this happens, obviously, there is no individual risk. Additionally, some times after an eventual pandemic swipes the planet, the risk will be close to zero again -- nobody's at risk of catching the Spanish Flu those days, haven't you noticed?

meetooAugust 24, 2007 11:14 PM

Antibozo,

I agree with Robert B, the data has been overblown.

If SHS really is as dangerous as the government, political organizations and charities claim, efforts to prevent it and contain it might be justified. But is it dangerous? We're bombarded by endless proclamations of its horrors, claims that get more fantastic with each passing year. These claims are usually accompanied by impressive sounding numbers. Are smokers really hurting every stranger in the vicinity? The answer to that question is obvious once you know the facts.

The Truth. The phrase sounds noble, solid and immutable. Once you arrive at The Truth your search is over. No one can argue with The Truth. Our courts claim it is their objective. Our journalists say they dig for it. Our religious leaders teach that it will set us free.

But people on both sides of every argument insist that they are telling The Truth. Politicians and other criminals swear to tell The Truth, the Whole Truth, and nothing but The Truth just before giving outrageously dishonest testimony. News programs blow up trucks with model rocket engines, then proudly display the journalism awards they've won for bringing us The Truth. Modern day snake-oil salesmen promise their high-priced, useless cures will restore our health. Thousands of backwater Bible thumpers assure their congregations they alone know The Truth, and that all their fellow thumpers, who also claim to know The Truth, are truly dupes of the devil.

Perhaps The Truth isn't as absolute as we assume it is. It's always colored by our perceptions. If I tell you my wife is beautiful, that's the truth. If you disagree with me, does that make you a liar? And if Truth is not absolute, what can we rely on?

The web site thetruth.com claims to offer The Truth about tobacco. They're really offering a smattering of facts mixed in with a great deal of propaganda and quite a few outright lies. We offer the antidote for their "truth." No, not our version of The Truth; that would make us as suspect as they are. We offer something that no imitation truth can stand up to. We offer Facts.

Facts are stubborn things. Facts are verifible. Facts don't change with the political mood. Once you know The Facts, The Truth becomes obvious.

http://www.davehitt.com/facts/index.html

antibozoAugust 25, 2007 1:19 AM

meetoo> Facts are stubborn things. Facts are verifible.

Facts are what you'll find at the URLs for the sample of studies I posted above. These facts are as verifiable and objective as facts can be, as they entirely based on quantities that are measured and analyzed in documented ways. To quote Robert B, "I challenge you to read the studies."

We can divert this discussion to the nature of truth, metaphysics, and religion all you want, but doing so won't make environmental tobacco smoke any less toxic, diminish the large volume of clear scientific evidence of this toxicity, or reduce the thickness of the smog of psychological denial surrounding the inconsiderate smokers who wish to rationalize their exposing others to their pathogens.

If you want to kill yourself with cigarettes, by all means, have at it, but don't expect the rest of us to inhale your musty halo of thanatos.

And if you really want to talk truth, I'd simply refer you to my various comments in this other thread for my point of view so we don't waste time:

http://www.schneier.com/blog/archives/2007/07/...

Porlock JuniorAugust 26, 2007 2:00 AM

In all this about smoke and all, has no one got around to answering bob's invocation of the old story about evil Rachel Carson killing babies with malaria? A good place to start would be
http://scienceblogs.com/deltoid/fastsearch?...

Short version: DDT, being well established as harmful to birds, was banned from agriculture, but not from mosquito control. However, it has gone out of use in many, or most, places, because of the rapid evolution of resistance among mosquitoes, which requires the use of other means of control. It is still in use where it's effective.

And the numbers given with the junkscience version of the story are, so appropriately to this particular comments thread, impossible. http://timpanogos.wordpress.com/2007/08/09/...

I would never suggest, however, that the reason for the ignoring of DDT resistance is its association with evil-lution.

Ed DarrellAugust 28, 2007 7:40 AM

Robert B:

Antibozo:

Just to clarify: are you saying that you believe your inherent civil liberties include the right to administer an airborne stimulant (nicotine) to passersby?

Studies available to Congress in the 1980s compared carbon monoxide in rooms with smokers to carbon monoxide levels on roads, levels deemed to be damaging to health, by the EPA and therefore requiring reductions. CO levels indoors often exceeded safety standards, and in a few cases exceeded the harmful levels by a lot. Between 1980 and 1985 Congress conducted extensive hearings on the health effects of tobacco smoke; you should be able to get copies of the hearings with studies, references, and cross-examination of the witnesses (on the Senate side, the Senate Committee on Labor and Human Resources); look for the hearings on cigarette labels, for example.

Nicotine a stimulant? Well, yeah -- but it's also a deadly poison. Before I staffed Congress, I worked in a botany lab at a university. We used nicotine spray to fumigate the greenhouses to kill all insects. Safety rules required the greenhouses to be vacated during fumigation and for some time thereafter. Every once in a while a bird or mouse would be trapped. Have you read the EPA pesticide registration for any nicotine-based insecticide? The skull and crossbones don't mean it was manufactured by pirates . . .

But of course, that's just the tobacco. Cigarettes are exempt from product labeling laws that require accurate listings of what is actually in the cigarette. The companies themselves purchase tons of odd leaves in order to mislead other companies about their formulae for cigarettes, but testimony in the 1980s established that most cigarettes have glycerine in them to make the tobacco appear moist and keep it clumping -- there was no information available about inhaling glycerine smoke -- and that's one that the companies will reveal. Tobacco companies also purchase deer tongue, a wild-growing herb that is sometimes called "dumb cane" for its property of paralyzing the tongue and other muscles of the mouth and throat, or sassafras, also known as file (as in gumbo), which is itself a rather potent human carcinogen -- for neither of these plants is there any testing to determine carcinogenic properties when smoked.

So, yeah, the statistical studies might be skewed; but don't for a moment think that all the skewing shows tobacco to be more dangerous that it is, especially since tobacco companies spent hundreds of millions of dollars over most of the 20th century to hide all information about harms.

Very few people are very good at accurately estimating all the risks they face everyday. A good general rule is to reduce the risks you face reasonably. Studies show that people who regularly buckle their seatbelts also slow down in school zones and drive closer (slower) to the speed limits than those who don't. Placebo effect? Perhaps. Who cares which statistical effect works when your life is spared?

AnonymousSeptember 17, 2007 6:07 PM

Second Hand Smoking:
frankly, I don't even care wether SHS is harmful or not. I just want the freedom to breathe good air. Even if SHS would turn out to be healthy I would not like it.

I could pee in your garden and argue that it is not harmful to you. I could argue that it is even good for the flowers. However, I am sure most people would still not appreciate me doing so.

Robert, if you really like to fight for freedom you would have to respect the freedom of choice of others to live smoke free - and stop smoking with other people around.

ElricSeptember 20, 2007 3:11 PM

Freedom to breath good air?

Sounds good to me, I like good air too. Sometimes it contains tobacco smoke. But if you want to ban being able to smoke in out door public spaces....
then stop driving your toxin spewing car with other people around.

Leave a comment

Allowed HTML: <a href="URL"> • <em> <cite> <i> • <strong> <b> • <sub> <sup> • <ul> <ol> <li> • <blockquote> <pre>

Photo of Bruce Schneier by Per Ervland.

Schneier on Security is a personal website. Opinions expressed are not necessarily those of Co3 Systems, Inc..