Nothing new but still worth repeating
From Rodger:
Here are the statistics from the National Safety Council: In the year 2000, firearms killed 600 Americans accidentally. That’s 600 out of nearly 280 million. Here are the other numbers of accidental fatalities for that year: autos, 43,000; falls, 16,200; poisons, 11,700; drowning, 3,900; ingestion of food or other object, 3,400. The only number of fatalities lower than accidental firearms deaths is that from poison gases – 400.
November 18th, 2003 at 2:27 pm
Well, that’s it!
I am officially calling today for the formation of Poison Gas Control International, the Poison Gas Policy Center, the Colition to Ban Poison Gases, the Poison Gas Injury Prevention Cooperative, and the Million Mol March
After all, the only reason anybody needs a poison gas is to kill another human being!
November 18th, 2003 at 3:28 pm
Ah, lies, damn lies and statistics! Gotta love it. Of course accidental gun deaths versus number of Americans is a totally meaningless comparison. A more meaningful comparison might be, how many Americans come in contact with a car every day, versus how many come in contact with a gun every day. And how many times per day? Then compare the deaths to THOSE numbers. But that wouldn’t look as good.
Also, I’m betting the number of intentional gun deaths is significantly higher than the number of intentional automotive deaths…
Mind you, I generally support gun rights, but I’m not afraid to call “bullshit” when someone skews statistics, even if it is in defense of a cause I generally support.
November 18th, 2003 at 3:34 pm
Well, you must have field day with the VPC and the Brady Campaign!
I would actually dare say people come into contact with guns more than they think. People who see me on the street don’t realize i have a 45 or a 9mm (depending on what i’m wearing). And criminals walk the streets with guns too. Plus, there are 100M gun owners. How many car owners? I dunno. It’s not a worthless statistic but it can be misconstrued.
November 19th, 2003 at 9:02 am
The question about contact was a fair one. About 40% (probably more) of American households have at least one gun in them, so you can assume that a little less than half of all Americans are in daily proximity to a firearm. Even with this in mind, you still have an impressively low accident rate.
If you have a pool in your yard, your kid (or your neighbor’s kid) is more likely to die in your pool than die because of a gun accident. If your kid rides a bike or has a canoe or a pair of roller blades, take them away! He’s safer if you take him shooting instead.
Fatal, childhood gun accidents are so rare that most doctors will go their entire lives without seeing one. Most people have no idea how uncommon these events are. Of course, one is too many, but a little perspective is appropriate.
November 19th, 2003 at 12:23 pm
I don’t disagree, but the linked story intentionally misuses and twists the numbers in invalid ways.
Never mind contact, how many people use a vehicle every day (either by driving one or riding in one) versus how many use a gun. I think you’ll find the former number to be exponentially higher. It’s completely comparing apples to oranges.
And again, what about the intentional deaths? Better than two thirds of the murders in the US (in 1999, at least) were shooting deaths. How many people were murdered with a car?
Not a fair comparison? Probably not; but neither is the link’s, and that’s my point.
I’m not trying to pass judgement on guns and gun ownerships. I’m merely trying to pass judgement on bullshit use of statistics.
November 19th, 2003 at 12:28 pm
Are there really that many gun owners? As I survey my circle of friends, very few of them own guns, and every single one of them owns at least one car.
I would even say that the 100 million gun owners statistic is a skewed one, since it doesn’t distinguish gun types. I would guess that somebody who keeps a loaded handgun in their nightstand is a lot more likely to be involved in an accidental gun death than someone who has a hunting rifle locked up in a case 11 months out of the year. Although a few idiot hunters blow themselves away each year… (No, I’m not saying hunters are idiots. I’m saying the ones who blow themselves away because they didn’t exercise gun safety are idiots.)
November 19th, 2003 at 1:41 pm
I’m merely trying to pass judgement on bullshit use of statistics.
Well, with respect, I don’t see that there is any misrepresentation here.
Say Uncle was clearly asserting that the risk of gun accidents is slight, especially when compared to other common hazards. He is correct.
I don’t understand why you keep bringing up intentional deaths, as they have nothing to do with accident rates. It also does not matter how often you shoot your gun compared to how much you drive. Sure, that would a valid point if we were asserting some sort of minute-to-minute comparison, but the figures he is quoting clearly reflect the average, overall risk.
Nobody talks about how harmful cigarettes are per minute or how dangerous a toy is per minute of play. The numbers presented above reflect the true, relative risks for people who drive an average amount, and for people who shoot an average amount. It’s a perfectly valid comparision (once you correct for the fact that more households have cars than guns, as we discussed).
And if you don’t believe the 40% figure, LOOK IT UP. It’s called “google”. Knock yourself out.
November 19th, 2003 at 4:37 pm
Mike:
I stand by my statement that the statistics are bullshit. They make unwarranted comparisons, comparing “apples to oranges,” as it were, and draws non-sequitur conclusions (or at least tries to lead you toward those conclusions).
The sole purpose of the inclusion of these statistics is to lead the reader toward this logic:
1. In 2000, only 600 people were killed in gun-related accidents.
2. That same year, nearly 72 times as many people were killed in car-related accidents.
3. Therefore, guns are safer than cars.
But those accidents involve innumerable variables that can not be simply ignored. Throwing around a lot of numbers and pretending that they can be compared at face value does not constitute a sound statistical foundation.
Let me give you an example: suppose in a given year, 1,000 people are killed in Honda Civics, and 100 people are killed in Volvos. Can I use these numbers to conclude that Volvos are safer than Civics? Not by themselves, no.
I must consider how many Civics there are on the road as compared to how many Volvos there are, and I must even consider the driving habits of people who drive one versus the other, before I can come to that conclusion.
So you see, without more data, the raw numbers (100 Volvo deaths versus 1,000 Civic deaths) can’t lead to any valid conclusions. And there, we’re even talking about very similar concepts: accidental automobile deaths, based on automobile type.
If you can’t see that different types of accidental deaths are even farther apart than that, and that comparing such numbers is even less valid, then I’m afraid anyone could use any old statistics they want, no matter how tenuous their relationship, to make any point they want, and you would buy it.
By the way, do a google of your own on the logical fallacies and you’ll see that the source post violates quite a few of them.
November 19th, 2003 at 9:50 pm
There is a very simple reason why presenting these statistics is valid. The latest ploy of the victim disarmament movement is to say they are striving for gun safety rather than actually banning guns. These numbers show that gun safety is not a serious issue.
I see no reason to doubt the 40% number for gun ownership. After all, just about everyone I know owns some. On the other hand, I’ll bet that almost no one who knows my wife is aware that she owns one.
November 19th, 2003 at 11:52 pm
I’ll gladly concede the 40% number. I wasn’t calling “bullshit” on that, I was merely surprised by it.
I understand the motivation behind the argument, but that doesn’t excuse the use of invalid comparisons, which is what the author has done. And even if it were valid, it still wouldn’t entirely address the safety issue, because the numbers only talk about accidental deaths, and ignore, say, injuries — also a part of safety.
Let me reiterate that I support gun ownership rights, and some in the forum seem to miss that. This just goes to show that I’m every bit as willing to call “bullshit” on faulty arguments that I’d otherwise support, as I am to call it on arguments I oppose.
I don’t want to take your guns away. I just want to take your bad math away. 🙂
November 20th, 2003 at 9:30 am
I just want to take your bad math away.
The math is quite valid. The only point that is in dispute here is which question we are trying to answer with it.
Your hypothetical example of Volvo vrs. Civic is a good one. We agree the comparison is meaningless unless we normalise the data to account for the different numbers of each car on the road. At that point, you still assert that the comparison is bad, because it did not take other variables (such as driving style) into account, whereas I would say that the comparision is quite valid at that point.
The difference boils down to what question you are trying to answer. “Do average Civic drivers crash more often than average Volvo drivers”? That’s a valid question, and it can be answered right now.
Of course, this may not be the fault of the car – further study, such as you suggested, might find that driving style or some other “confounding variable” is to blame. That would help answer the question “Which car is safer in the hands of a given driver”? As I’m sure you see, this is a very different question, and it requires different data to answer it.
My point, all along, is that the first sort of question is a perfectly valid one, and not an example of the misuse of statistics. Furthermore, it’s been answered in a perfectly valid way. To be more specific, the question we’ve been discussing al lthis time is this one:
“How does the overall risk of gun accident compare to the overall risk of car accident”?
That’s a valid question.
I suspect the question you want to ask is this:
“For a given, average person who both drives and owns a gun, which is more likely to be involved in his accidental death, the car or the gun?
These are very different questions, and may yield different results. However, given the huge disparity in the numbers, I’d bet good money that even if you could isolate all the confounding variables that you could imagine, you’d still see the results come out the same. I mean, seriously, can you imagine some hidden factor between these questions that would account for something like a 35-fold difference? I can’t.
(The only way I suppose you could get away with it would be to use the “minute-by-minute” approach, which is itself deceptive. I could use that method to show that sex is more dangerous to your health than smoking, on a minute-by-minute basis)!
November 20th, 2003 at 1:14 pm
Mike:
My complaint is less complicated than you seem to think. My point, all along, is that the numbers aren’t normalized, and that can’t really tell us anything. As a side point, they only concern themselves with deaths and ignore injuries. Safety’s about more than whether you survive.
The only thing that the provided statistics do tell you is that you are statistically more likely to be killed by a car than by a gun. To which I say, “No shit, Sherlock!” You’re also statistically more likely to be killed by a car than, say, by somebody crashing a jumbo jet into the building you happen to be in at the moment.
It does not tell you anything about the relative safety of the things being compared, but he purports to do exactly that.
By the way, to extend his (and your) logic, poison gas is even safer than guns! So you really shouldn’t mind keeping poison gas around the house. The numbers don’t lie!
To be more specific, the question we’ve been discussing al lthis time is this one:
“How does the overall risk of gun accident compare to the overall risk of car accident”?
That’s a valid question.
Too bad the numbers still don’t answer it. They tell you the number of deaths, not the number of accidents. The number of related deaths is probably somehow correlated with the number of accidents, but I doubt that they’re correlated at the same rate across the board. (For example, what percentage of car accidents lead to death versus what percentage of gun accidents lead to death?) So even that question isn’t answered by these numbers.
But if you don’t mind comparing apples with freight trains, then there’s probably nothing I can do to convince you that his statistics are quite meaningless.
November 20th, 2003 at 1:18 pm
Actually, one of my own statements above isn’t entirely true:
Should read:
The only thing that the provided statistics do tell you is that you are statistically more likely to be accidentally killed by a car than by a gun. To which I say, “No shit, Sherlock!” You’re also statistically more likely to be accidentally killed by a car than, say, by somebody crashing a jumbo jet into the building you happen to be in at the moment.
I should also clarify that the rate of accidents as compared to usage is only one dimension of the safety equation. The other, perhaps equally important dimension is the frequency, severity and nature of injury in the event that an accident occurs.
November 20th, 2003 at 2:04 pm
Shame, shame, shame on me! I’m always chiding Uncle for not doing his homework, and I didn’t do mine. It always struck me as odd that Rodger cited those numbers without linking them, but I didn’t verify them myself.
So I went out to look at the actual NSC numbers for 2000, and whaddya know? His numbers are even more bullshit than I ever imagined!
Quoth Rodger:
Oh, really? According to the numbers I see, there were 776 accidental deaths owing to firearms, meaning he underrepresented that number by more than 29%. Meanwhile, cars accounted for 14,813 deaths (18,450 if you include pickups, vans, SUVs, and “heavy transport vehicles”), meaning that he over-represented this number by at least 233%, and possibly by as much as 290%.
Even if we accepted his numbers, the logic was bullshit, but now we see that even the numbers are bullshit.
November 20th, 2003 at 2:14 pm
OK, this is the last I’ll post on this.
If you think that cars are not really more dangerous than guns, then please, demonstrate that for us. Show me data that support your assertion. You may include deaths only, or injures and deaths combined, whichever you prefer.
November 20th, 2003 at 2:21 pm
Well, the CDC jives with the 776 number:
http://www.cdc.gov/nchs/data/nvsr/nvsr50/nvsr50_15.pdf go to page 69.
And on page 30, the 43,000 number jives.
November 20th, 2003 at 2:46 pm
Mike
OK, this is the last I’ll post on this.
If you think that cars are not really more dangerous than guns, then please, demonstrate that for us. Show me data that support your assertion.
Huh? The hell you say? First, when have I ever said that cars aren’t more dangerous than guns? I’ve never said any such thing! All I have said is that Rodger’s numbers fail to prove that they are.
But when did the burden of proof shift to me? Rodger has made an argument which does not withstand logical and statistical scrutiny. Me saying that he hasn’t sufficiently supported his argument does not, in any way, obligate me to disprove his argument or to prove the opposite. Except maybe in your world.
My only “assertion” is that his statistics are bullshit, and I’ve provided evidence (both anecdotal and hard) to back that up.
Uncle:
I’d be curious to know what the CDC classifies as a “motor vehicle,” because those numbers are difficult to sqaure with the NSC’s numbers. Certainly, if the NSC’s numbers are accurate, the CDC must be using a much broader definition of “motor vehicles” than just cars.
November 18th, 2003 at 10:05 pm
Lies, damn lies, and statistics.
There’s a nice little point-counterpoint over at Say Uncle today. It’s gun stuff, and it’s civil! Wow….