LESSWRONG
LW

Comment Permalink

Thank you for your well-researched response.

it might be the case that the methodology in that study is significantly more reliable than that of other studies on this subject. If you are aware of an argument to that effect, I'd be interested to hear it.

As you pointed out, the study you're citing is gated, so I can't really evaluate it. However, based on the portions you quoted, I don't see any reason to believe that either study is more or less methodologically sound than the other. However, I also don't see any reason to believe that the two studies actually contradict each other, despite first appearances. The study I cited evaluates national-level crime stats, whereas yours evaluates household-level victimization stats. Both studies' conclusions could easily be true simultaneously.

As noted in your citation:

None of the studies can prove causation and none have completely eliminated the possibility that the association might be entirely due to reverse causation

My intuition is that reverse causation is at play. That may sound like a cop-out, but keep in mind, I also argued reverse causation for my own citation, despite that being against my political interests, so I'm not just privileging my own position here. In the case of your citation, the intuitive case for reverse causation sounds even more convincing to me: people who live in high-crime neighborhoods are much more likely to decide to buy a firearm, in response to their local crime rates - I can attest to this from personal anecdotal experience.

Since prevalence of violent crime varies neighborhood-by-neighborhood, while gun control laws typically vary only nation-by-nation, there could very well be a positive correlation at the neighborhood (or household) level, while simultaneously having a negative correlation at the national level... especially in light of the fact that your study's abstract concedes that its results are dominated by a single nation (the US).

The (informal) null hypothesis - that gun control laws have no significant causal effect, either positive or negative, on violent crime rates, would seem to explain both studies' results equally well, whereas assuming the existence of (non-reverse) causation, either positive or negative, would require either one or the other of the studies' conclusions to be wrong somehow. Parsimony would therefore seem to recommend this hypothesis.

See in context

17 Gun Control: How would we know?

by rlpowell

20th Dec 2012

2 min read

169

17

I don't know how to keep this topic away from http://lesswrong.com/lw/gw/politics_is_the_mindkiller/ , so I'm just going to exhort everyone to try to keep this about rationality and not about politics as usual. I myself have strong opinions here, which I'm deliberately squelching.

So I got to thinking about the issue of gun control in the wake of a recent school shooting in the US, specifically from the POV of minimizing presumed-innocents getting randomly shot. Please limit discussion to that *specific* issue, or we'll be here all year.

My question is not so much "Is strict gun control or lots of guns better for us [in the sole context of minimizing presumed-innocents getting randomly shot]?", although I'm certainly interested in knowing the answer to that, but I think if that was answerable we as a culture wouldn't still be arguing about it.

Let's try a different question, though: how would we know?

That is, what non-magical statistical evidence could someone give that would actually settle the question reasonably well (let's say, at about the same level as "smoking causes cancer", or so)?

As a first pass I looked at http://en.wikipedia.org/wiki/List_of_countries_by_intentional_homicide_rate and http://en.wikipedia.org/wiki/List_of_countries_by_firearm-related_death_rate and I noted that the US, which is famously kind of all about the guns, has significantly higher rates than other first-world countries. I had gone into this with a deliberate desire to win, in the less wrong sense, so I accepted that this strongly speaks against my personal beliefs (my default stance is that all teachers should have concealed carry permits and mandatory shooting range time requirements), and was about to update (well, utterly obliterate) those beliefs, when I went "Now, hold on. In the context of first world countries, the US has relatively lax gun control, and we seem to rather enjoy killing each other. How do I know those are causally related, though? Is it not just as likely that, for example, we have all the homicidally crazy people, and that that leads to both of those things? It doesn't seem to be the case that, say, in the UK, you have large-scale secret hoarding of guns; if that was the case, they'd be closer to use in gun-related homicides, I would think. But just because it didn't happen in the UK doesn't mean it wouldn't happen here."

At that point I realized that I don't know, even in theory, how to tell what the answer to my question is, or what evidence would be strong evidence for one position or the other. I am not strong enough as a rationalist or a statistician.

So, I thought I'd ask LW, which is full of people better at those things than I am. :)

Have at.

-Robin

Personal Blog

17

Gun Control: How would we know?

New Comment

169 comments, sorted by

top scoring

Click to highlight new comments since: Today at 6:11 AM

Some comments are truncated due to high volume. (⌘F to expand all)Change truncation settings

[-]Emile12y390

if that was answerable we as a culture wouldn't still be arguing about it.

Who is "we", white man? ;)

Here in France there seems to be very little public debate around firearm legislation - there probably is some, somewhere, but it gets very little media attention.

The existence of public debate around an issue doesn't seem to be very strong evidence on whether the question can be answered conclusively given the available evidence, since public debate around issues varies from country to country, and in some places there are public debates around issues where anybody with half a brain should be able to tell which side is right. Heck, sometimes there's a lack of public debate because everybody agrees on the wrong conclusion.

3ewbrownv12y

Agreed. Presence or absence of debate on an issue gives information about a nation's culture, but very little about how hard it is to discover the facts of the matter. This is especially true in matters of social science, where the available evidence is never going to be strong enough to convince someone who has already made up his mind.

0FiftyTwo12y

Likewise in the UK, which is otherwise culturally very similar to the US. Does it provide any interesting evidence that in countries with policy P (available guns) there is a lot of public debate and in countries with policy ~P (restricted guns) there is little to no debate? It would seem to imply that the supposed harms of ~P are either non-existent or insufficiently obvious to reach public debate. (Same appears to apply with public health care, and I'm sure there are other examples for different pairs of countries).

4Sabiola12y

Here in the Netherlands guns are severely controlled, and the only times gun control is discussed at all (after a shooting makes the news), people just say 'Those crazy Americans - of course guns should be controlled!', and go on complaining about the weather. I wonder if gun control is a hot topic Switzerland, where many people have guns.

1kodos9612y

This kind of argument is what I like to call "motivated majoritarianism". You're essentially assuming that public opinion is rational - something that most LWers would be loath to do in the general case.

1FiftyTwo12y

I don't assume public opinion is rational, but that it is sufficiently predictable that with an issue where in one case it is discussed very heavily it would be in a contrary case if the benefits were comparable. When what should be a double sided issue is not being discussed you require an explanation for that behaviour.

0Peterdjones12y

Or that they are localised. The pro-gun arguments in the US boil down to 1) The constitution -- doesn't apply elesewhere 2) A multitude of armed criminals (ditto).

0A1987dM12y

Maybe stricter gun control would be beneficial in certain countries and harmful in others, depending on all kinds of stuff.

[-]AlexMennen12y360

(my default stance is that all teachers should have concealed carry permits and mandatory shooting range time requirements)

Let's assume that your suggested policy would bring school shootings from about the rate they're at now to 0. I can't imagine the benefit would be much better than that, and it would probably be a lot worse. Wikipedia says that there have been 38 school shooting deaths this year (not including the suicides, and including the recent attack, making it much higher than other recent years). According to this, there are about 3 million public school teachers in the US and they make about $50,000 per year each, so their value of time is probably somewhere around $30/hour, so it would cost about $100 million per year to require all of them to spend an hour per year on the shooting range. If that saves about 40 lives per year, that works out to $25 million per life (Edit: oops, no it doesn't). None of the estimates on wikipedia suggest that lives should be valued at more than $10 million per life. And I haven't even mentioned the costs of equipping the teachers with guns, so the actual cost of the policy is probably much higher. So mandatory firing range time for al... (read more)

[-]Alsadius12y140

Your numbers don't add. $100 million/year divided by 40 deaths/year is $2.5 million per life, which is well below the accepted value of a life.

2AlexMennen12y

Oops. You're right. But since I was trying to make the number as low as possible so no one could claim it would be lower, it's still almost certainly well over $10 million/life. If we look at the recent few years, there's been about 10 per year, so if we're still being pessimistic, we have to assume that's a trend and that there'd be about 100 fatalities per year otherwise, putting it at $10 million/life, still just over the value of statistical life. And still almost certainly an underestimate since 1 hour/year of teacher training is a massive underestimate of the cost, there's no way it would send school shooting fatalities to 0, etc. But it's closer than I expected it to be, though.

0ModusPonies12y

$2.5 million is a thousand times greater than the marginal cost of saving a life via effective altruism.

[-]A1987dM12y130

But the lives you'd save via effective altruism are not American lives! ;-)

0Alsadius12y

Yes, but nobody expects government action to be effective.

[-]FiftyTwo12y130

We can also look at comparative advantage. If we are dedicating this length of time would ti eb better spent on something else, say teaching them all first aid? I suspect there are significantly more deaths from accidents than shootings.

[-]Zack_M_Davis12y100

(Alex scooped me on the obvious "do a back-of-the-envelope calculation" point; below is the calculation I was going to include in the comment I was drafting)

There are about 7.2 million teachers in the United States. Suppose firearms training takes ten hours per teacher per year, and suppose we value that time at 15 $/hr; that's already more than a billion dollars. How many people die in school shootings in a typical year? Glancing at Wikipedia's list of U.S. school shootings in the decade 2000-2009, I count 67 fatalities, or about 7 per year. But it doesn't seem reasonable to spend $140,000,000 to save a life [...]

1Randy_M12y

Based on those kinds of calculations, teacher in service meetings or parent teacher conferences cost more lives than this school shooting. eta: Not that there isn't a valid point that arming every single teacher, even if effective, would probably be overkill. Having one per x number of students (or x square feet of campus size) would probably be just as effective and not actually require training, as there are likely a few already there. And the guns wouldn't necessarily need to be on their person or in their classroom--just somewhere closer than the nearest police station.

2gwern12y

Only if you quietly assume it has nothing to do with the important task of teaching.

1Randy_M12y

Right, that was the implication. I don't think they tend to do much to improve the quality of babysitting. (Okay, somewhat harsh, but I was a teacher.) but surely even if they did, it wouldn't be enough to justify the cost in strict utilitarian calculations like the above?

3gwern12y

I don't know, why wouldn't it be? A high-tech wealthy economy depends on education, with all the direct & indirect returns implied.

1A1987dM12y

Not to mention that certain people might be good at teaching but bad at shooting. (Wild-ass speculation here -- I've never fired a real gun and I have no idea if it's something almost everyone could learn to do decently.)

7NancyLebovitz12y

My impression is that almost anyone can learn to do target range shooting decently, though even then you'd run up against disability issues if you made that level of skill a requirement for teachers. The hard part is staying calm enough to do something useful (or perhaps anything at all) in a fast-moving violent situation. This can take a lot of training. On the other hand, I haven't heard about any of the teachers freezing at Sandy Hook. On yet another hand, most people have powerful inhibitions against killing, so that might be harder to train than protecting children.

1Nornagest12y

It is very hard to teach people to kill -- a fact that's largely responsible for the lopsided casualty ratios in engagements between well-trained and poorly trained armies. Most of the research that has been done on this is in the context of police and military training, though, where the main motivating factor is not letting your buddies down; I'd expect the psychology to be somewhat different in the context of defending children. Given that teachers self-select for very different psychology than cops or soldiers, though, I'm not sure which way the statistics would end up running.

5Stefie_K12y

There are further implications along these lines, too. It's isn't just ability, but willingness: at least some prospective teachers would probably be put off by the prospect of being required to be armed in the classroom. Not that the job market for teachers isn't glutted, right now, but is "willingness to carry a gun and shoot to kill" really something that we want to select for, in teachers? It would compete with the ability to teach well in determining who actually teaches our children.

[-][anonymous]12y240

If gun control arguments make me want to shoot myself, does that just prove their point? by Yvain

I have tried to be good.
I have tried not to talk about politics on Facebook, because that's not the place for it, and it only annoys people, and it's not what people want to hear about right after a terrible disaster.
No one else has tried this. I don't think people who post about politics on Facebook all the time realize that everyone else who agrees with them is also posting about politics on Facebook all the time, and so every day I have to scroll through half a dozen image macros making fun of how stupid anyone who doesn't want immediate gun control is, or catchy anti-NRA slogans. The day after the tragedy, there was almost nothing else in my entire newsfeed.
The posts are never "I think we need more gun control". It's always "Anyone who doesn't want gun control has been brainwashed by the NRA and thinks school shootings are great." I am constantly amazed by how small a buffer the average person has between "I don't believe X" and "Believing X is irredeemably evil and we must mock and shame it until the very possibility of expressing it is beyond t

... (read more)

0rlpowell12y

"No one else has tried this." -- I have, actually, which is why this post is here. :D Thanks for the great link, that's the sort of thing I was wanting to see. -Robin

-2prase12y

And the explicit point in relation to rlpowell's post? Are you objecting to discussing gun control se early after a rare event associated with guns which could skew people's thoughts via availability heuristics, or are you attempting to counter these biases in the conversation, or something else?

0rlpowell12y

I don't know why Konkvistador posted so much of Yvain's article, or highlighted the particular parts of it that ey did, but the article itself goes into the research on this topic in some detail, which certainly hepls. -Robin

3prase12y

Yvain's article is great (which is usual for his articles), but its topic is political debate about gun control rather than gun control per se. The reason why I am asking is that I recall Konkvistador asking people not to post about politics around the time of the last US elections, so in my model of Konkvistador it would make sense if he asked people not to post about gun control around the time of the last US school shooting. On the other hand, his comment isn't exactly what would be expected if that was really his intention, so I am a bit confused.

[-][anonymous]12y190

With this kind of question that borders on political/identity issues, they very first thing you do is build your lines of retreat: If guns cause death, what should you do? If guns don't cause death, what should you do? If guns reduce death, what should you do? You need satisfactory non-straw answers for all of those cases before you are qualified to look at the facts. After you have those answers, you will find it much easier to be neutral.

I think you'd have to notice the (anti)correllation between gun control and violent death across countries/states/areas, if it exists, then look at the three hypotheses: guns cause death, death causes guns, something causes both. Try to eliminate hypotheses.

You could note that Canada has more guns per capita and less murder. You could control for the effect of legislation by finding different groups of people who kill each other with guns or don't who live in the same legislative areas. You could find a culturally homogenous group who are split by arbitrary political lines (i.e. toronto/detroit/buffalo/etc, that country that is split in two in the carribean). etc.

[-]Jabberslythe12y160

Canada actually has fewer guns per capita. The US is definitely topping that list.

[-][anonymous]12y140

Someone has been feeding me lies!

6Jabberslythe12y

I believed it as well until recently. I first remember hearing it in "Bowling For Columbine" so maybe Michael Moore is the culprit. American having more guns per capita definably fits better with my my background beliefs of those two countries.

6satt12y

Or did you? I Ctrl-Fed through a few versions of the English subtitles available for Bowling for Columbine to try to find a claim that Canada has more guns per capita than the US. I turned up a blank (although subtitles don't capture everything, of course). The most relevant bits I found were subtitles 1580 to 1583: That implies about 23 guns per 100 Canadians, which is if anything an underestimate.

0Jabberslythe12y

Whoops, I guess my memory was being confabulatory.

0[anonymous]12y

Not long ago, I read (but I forgot where, and now I can't find it) that in the US many people have lots of guns, instead of just one, so that would skew the statistics. Numbers pulled out of my posterior: If every household in Switzerland has a gun, and in the US only half of the households has a gun, but half of those who do have 3 or more, then the number of guns per capita would still be higher in the US than in Switzerland.

0A1987dM12y

Yeah, I think I got that from there too.

3JonathanLivengood12y

A more interesting number for the gun control debate is the percentage of households with guns. That number in the U.S. has been declining -- pdf, but it is still very high in comparison with other developed nations. However, exact comparisons of gun ownership rates internationally are tricky. The data is often sparse or non-uniform in the way it is collected. The most consistent comparisons I could find -- and I'd love to see more recent data -- were from the 1989 and 1992 International Crime Surveys. The numbers are reported in this paper on gun ownership, homicide, and suicide -- pdf. These data are old, but in 1989, about 48% of U.S. households had a firearm of some kind, compared with 29% of Canadian households. However, the numbers for handguns specifically were very different. In 1989, only 5% of Canadian households had a handgun, compared with 28% of U.S. households.

2A1987dM12y

By a factor of three. Whereas the number of firearm-related deaths... [looks it up] I'm surprised. Turns out that the average gun in Canada kills more people than the average gun in the US, though most of that is suicides. The average gun in the US does kill more people other than its user than the average gun in Canada, but it's within a factor of two.

2Alsadius12y

If you look at long guns alone, or actual gun owners(instead of guns owned), the numbers are closer. The US is still definitely in the lead, though.

4taelor12y

Point of order: I'm assuming that by "that country that is split in two in the carribean" you are referring to the island of Hispaniola, which is divided between the nations of Haiti and the Dominican Republic, which are not especially culturally homogenous (if your refering to a different carribean island, then ignore this).

1[anonymous]12y

I was referring to that one. I expected them to be culturally similar because that border is awful straight, but if they aren't, they aren't. Do you know more about what the differences are and where they came from? Is it differences in political history? Different groups of people? EDIT: Toronto and Detroit aren't all that similar either. I wonder where you could find a good case where legislation is the only difference between two populations.

[-]Alsadius12y100

Haiti and the Dominican have remarkably different political histories - Haiti has a lot more awful dictators in its past. They're poor comparisons. Toronto and Detroit are probably even worse - Detroit is an industrial centre whose industry rotted out around the same time as it was levelled(metaphorically) by race riots and people fleeing from them. Toronto is a financial and cultural centre whose primary challenge is that they're short a freeway or two and the commute times are bad. If you want to compare Toronto to anything, NYC or maybe Chicago are the obvious candidates.

As for a close cross-border comparison, I propose Seattle and Vancouver. They're close, similarly sized, and culturally close. You can also look at smaller examples(Sault Ste Marie ON/MI, Niagara Falls ON/NY, etc.), but it's harder to get good data from those due to scale, even though they are even closer in geographic and cultural terms.

0taelor12y

North and South Korea just after the split is probably closest to what your looking for, though currently, they have 50 years of differing political histories separating them. Still probably better than most examples.

0rlpowell12y

In my limited-ish experience, some Canadian border towns (Niagra Falls, in particular) get pretty close, but you'll get lots of people on both sides that concentrate on their national identity.

0taelor12y

Both countries share a cultural inheritance from the native Taino people, but how much of that was able to escape post-contact ethnocide is a matter of debate. The two countires were united under first Spanish and then French colonial rule, but following the end of colonialism, the mainly Spanish-speaking east engaged in a hard fought war of independance from the French-speaking west. The "awful straight" border you noticed was the end result of this war.

0rlpowell12y

I don't have clear lines of retreat for the simple reason that to answer what I would do in each of those cases requires also knowing what sorts of actions make things better in each case. I mean, I can say something generic like "increase or decrease the availability of guns in linear proportion to how much they help", but what actually decreases availability of guns, without having terrible side effects? Like, does gun control as we currently understand it lead to only crazy/criminal/insane people owning guns?, because that seems suboptimal. Having said that, I feel pretty confident that I'm willing to follow the data here; I think I've dismantled my ego need to support my historical position pretty well over the last few days.

-3printing-spoon12y

Lines of retreat are for offering to other people during arguments. I think I can trust myself to be neutral.

[-]kodos9612y130

No. Allowing yourself a line of retreat helps disincentivize the less-rational parts of your brain from stubbornly insisting on continuing to defend a proposition after it is no longer viable.

-6printing-spoon12y

8[anonymous]12y

Be sure to leave a line of retreat from that option too.

[-]Oscar_Cunningham12y150

Ban guns in a randomly chosen* selection of 25 states and see if fewer people get shot in those states.

*"randomly chosen" is important to avoid effects from regression to the mean etc.

9RobertLumley12y

The borders of gun states vs non gun states would be interesting.

5Eugine_Nier12y

Not necessarily since it's easy to move guns across them.

0Alsadius12y

Probably not what Robert meant, but I'd love to watch trafficking patterns across those borders.

2jefftk12y

Though "get shot" seems less what we're trying to minimize than "get killed".

[-]Desrtopa12y140

(my default stance is that all teachers should have concealed carry permits and mandatory shooting range time requirements)

Even assuming that this eliminates all school shootings and doesn't result in even a single child getting their hands on a teacher's gun which isn't properly attended, or a teacher lacking in self control using one irresponsibly, I think this would probably still be a poor use of time and money with respect to lives saved relative to other equipment and training (emergency medical, for instance,) that teachers could receive. It sounds like much more a response to Bad Guy Bias than lives-saved maximizing.

2Alsadius12y

Presumably those teachers would be armed in situations outside school. Throwing a few million people who carry concealed on a daily basis onto the streets would probably have some major effects outside of schools. (Whether net-positive or net-negative is left as an exercise for the reader)

4A1987dM12y

In the standard I-have-no-idea-either sarcastic sense?

4Alsadius12y

Partially that. Partially, I don't want to presume to answer in a parenthetical aside a question that this thread is about figuring out how to answer in the first place. I figure LW is about the worst possible place to vomit cached thoughts on controversial issues onto the page. (Also, I get a kick out of playing the neutral moderator in Internet debate - it always cracks me up how much people flip out on you when you make no statements of opinion whatsoever and merely posit viewpoints and mock bad arguments. Abortion is especially good for this.)

1A1987dM12y

I used to do that on WP:Requested moves.

3Desrtopa12y

Some probably would, but many would probably only carry the guns at all under sufferance, so it's worth considering also what impact it would have on the population that's interested in teaching.

4Alsadius12y

True, adding an irrelevant-but-mandatory criteria for taking up a job is bound to have some effect on the population in that job. I wonder what the result would be on the perceived leftist bias of the education system if you forced the system into 100% gun ownership(which is, these days, one of the strongest Republican indicators).

0Randy_M12y

Is it? I've read recently that 40% of gun owners are Democrats, although I couldn't remember where at the moment. I could think of more reliable indicators. Well, actually, I thought I could, but in fact I could think of more reliable Democrat indicators but I'm not sure about R ones.

0Alsadius12y

I'm getting my data from http://fivethirtyeight.blogs.nytimes.com/2012/12/18/in-gun-ownership-statistics-partisan-divide-is-sharp/ Also, I meant demographic indicators - if opinions count, "I am a (Republican/Democrat)" seems the most accurate.

0Desrtopa12y

Advocacy of teaching Creationism or Intelligent Design in schools? Opposition to both gay marriage and civil partnerships?

0Randy_M12y

Yes, probably. Homeschooling, perhaps, as well, though maybe only if you stick to a D/R dichotomy.

0Nornagest12y

I was thinking about this a couple days ago. It seems to me that stationing one or two police at schools would be more effective than this idea if we're interested in taking an active defense approach to stopping school shootings or minimizing their harm, though I hasten to add that they're a tiny fraction of total homicide and probably don't deserve this kind of attention. My thinking is that uniformed officials with the weight of authority behind them would probably have a more salient deterrent effect than whatever armed schoolteachers would imply, and also that forging non-confrontational links between police forces and civilians would have substantial positive externalities. Though this latter would depend greatly on style; a scare-'em-straight approach might backfire, and I'm almost sure that using cops purely as glorified security guards would. Arming and training teachers might put shooters down faster, but I can't see much deterrence or any substantial externalities, and it would be an expensive program. Though on the other hand I don't see many negative externalities either; last time I looked at data on shootings (accident and murder inclusive) attending trained and licensed bearers of firearms, the rates turned out to be quite low.

0Desrtopa12y

Some schools already do this. My high school, for instance. But that was a school with a particularly low level of threat to begin with. Not that there was never any crime for him to deal with even in a well off suburban school; the worst case I ever heard about him having to deal with was a stolen laptop.

[-]NancyLebovitz12y120

A impressive essay. It's not exactly on topic, but I think it will appeal to LessWrongers as a clean presentation on a mind-killing topic. It's by a pro-gun leftist, and it's the first explanation of the definition of assault weapons that I've been able to focus on. The author is remarkably clear about the fact that other people don't start out by sharing his knowledge. There are no insults, and there are explanatory pictures.

0kodos9612y

Excellent essay, thanks for the link. Since I was already familiar with most of the facts, this was the part that stood out to me the most: So my main take away from the essay was "damn, I should start stockpiling assaulty-looking guns as an investment, since they're likely going to be re-banned now". Does that make me a horrible person?

0NancyLebovitz12y

I don't think it makes you a horrible person, though it does seem to be surprisingly difficult to make a lot of money fast by making people's lives better. Perhaps it's not surprising that there's a lot of money in supernormal stimuli. I've been reading *Antifragile", so.... what's the downside? Would you be better off just getting an option on a bunch of scary-looking weapons?

1Alsadius12y

Why do you assume that this doesn't make people's lives better? You know perfectly well that the vast majority of those assaulty-looking guns are just going to wind up with collectors who already own enough firepower to conquer Hawaii. But owning banned stuff makes them happy, and what's wrong with that?

1NancyLebovitz12y

Ruby Ridge. Even occasional enforcement can have high costs.

0Alsadius12y

What does Ruby Ridge have to do with turning a profit on ban arbitrage?

0NancyLebovitz12y

I was disagreeing with your idea that the ban makes people's lives better.

1kodos9612y

The two of you seem to be talking past each other, so I think it would be useful if you both stepped back and stated in specific, concrete terms, what position you're actually arguing for.

0Alsadius12y

I did not mean to assert that. I meant to assert that being able to buy guns despite the ban makes people's lives better if you take the ban as an exogenous fact. I can understand the confusion though, my original comment was somewhat ambiguous.

[-]Swimmy12y120

There are two obvious effects (guns are more deadly than other weapons, but guns are also a deterrent) and it is not clear which is stronger. It's one of those issues where natural experiments or instrumental variables are our best bet for gaining knowledge, and of course anyone with a stat background will know the troubles with those techniques.

That said, there are studies using those techniques and they are better than a cursory glance at gun laws and homicide rates by country (or by state). And, to my understanding, the results of those studies are resoundly mixed. Some of these are quite controversial, but we're talking about tricky statistical techniques surrounding an emotional political issue, so controversy will abound even if the results are sound.

My takeaway is that this is not an issue worth getting very passionate about one way or the other. Your knowledge should drive your emotions, and if you don't know what effect is strongest, then you should save your emotional energy for more clear-cut or important causes.

If anyone knows of any very elegant studies, please correct me. Obviously I haven't read the whole literature.

2rlpowell12y

Oh, I like this. I like this a lot. The underlying attitude, I mean. I'm going see if I can't extrapolate a general policy from this, actually. Something like: "In a world where there still exist children that live (or, more likely, die) on garbage heaps, the fact that we're still arguing about [whatever issue] implies to me that it's not a low hanging fruit, and we should just go work on those instead."

[-]kodos9612y90

While it's almost certainly impossible to answer this question to anywhere near the level of "smoking causes cancer", it's surely possible to get much closer than just comparing those two statistics you cited. One of the best attempts I've seen (and not just because it happens to support my position, I swear) is this study from the Harvard Journal of Law & Public Policy.

It attempts to find a statistically rigourous correlation between gun ownership rates and murder/suicide rates - importantly, not JUST gun-related deaths, but murder/suicide r... (read more)

4pragmatist12y

My go-to for getting a sense of the empirical literature on some issue is to look for review articles that aggregate the results of multiple studies. I found this review, which suggests that there is a correlation between gun prevalence and homicide (and not just firearm-related homicide). The article is gated, sorry, but here's the abstract: And here's the conclusion: I haven't read through the study you linked, and it might be the case that the methodology in that study is significantly more reliable than that of other studies on this subject. If you are aware of an argument to that effect, I'd be interested to hear it.

4kodos9612y

Thank you for your well-researched response. As you pointed out, the study you're citing is gated, so I can't really evaluate it. However, based on the portions you quoted, I don't see any reason to believe that either study is more or less methodologically sound than the other. However, I also don't see any reason to believe that the two studies actually contradict each other, despite first appearances. The study I cited evaluates national-level crime stats, whereas yours evaluates household-level victimization stats. Both studies' conclusions could easily be true simultaneously. As noted in your citation: My intuition is that reverse causation is at play. That may sound like a cop-out, but keep in mind, I also argued reverse causation for my own citation, despite that being against my political interests, so I'm not just privileging my own position here. In the case of your citation, the intuitive case for reverse causation sounds even more convincing to me: people who live in high-crime neighborhoods are much more likely to decide to buy a firearm, in response to their local crime rates - I can attest to this from personal anecdotal experience. Since prevalence of violent crime varies neighborhood-by-neighborhood, while gun control laws typically vary only nation-by-nation, there could very well be a positive correlation at the neighborhood (or household) level, while simultaneously having a negative correlation at the national level... especially in light of the fact that your study's abstract concedes that its results are dominated by a single nation (the US). The (informal) null hypothesis - that gun control laws have no significant causal effect, either positive or negative, on violent crime rates, would seem to explain both studies' results equally well, whereas assuming the existence of (non-reverse) causation, either positive or negative, would require either one or the other of the studies' conclusions to be wrong somehow. Parsimony would therefore s

3Jayson_Virissimo12y

What are some of the non-best attempts you've seen?

1kodos9612y

Well, back of the napkin style comparisons like the one in the OP. And the press releases put out by the lobbying groups (on both sides of the issue).

1Randy_M12y

I agree that it is important to look at homicide rates rather than rates of gun deaths. (Relevant) arguments against gun restriction are basically that, first, it isn't practical to significantly reduce the amount of gun ownership amongst the prone-to-impulsive-violence population, and, second, that if you did other deadly implements abound, especially ones that magnify strength disparities that likely correlate with being violence prone. Looking at just gun death rates after restriction can argue against the first (given other variables controlled) but murder rates are the more important stat and could update both hypothesis one way or another.

0drnickbone12y

This is interesting, but correlations of such generic variables without a clear causal story to explain them are a bit suspicious: there are so many confounding factors. How does gun possession reduce homicide from other causes for instance (Deterrent effect? Dubious since most homicides are heat-of-the-moment crimes of passion). How does gun ownership reduce suicides? The original poster asked for evidence on a par with "smoking causes cancer", and this Harvard study seems a lot like comparing smoking rates in a country against overall death rates in a country. (There might well be a negative correlation there too e.g. if countries with more smokers are generally more developed and so have lower death rates overall).

3kodos9612y

I concur. As I said, I find these results counterintuitive, despite the fact that they superficially seem to support my political position. They seem to support my position TOO WELL, and that makes me quite suspicious of them. If I had to take a guess, I'd say that in reality, gun control laws have very little effect, in either direction, on crime rates, and that the causation behind this correlation, if there is any at all, runs in the opposite direction as would be implied by the naive reading: countries which have high crime rates tend to pass strict gun control laws as a reaction to those high crime rates, while countries with low crime rates never bother passing gun control laws, since they don't see the need. In other words, I think it's more likely that high crime rates cause strict gun control laws, rather than strict gun control laws causing high crime rates.

[-]Kingoftheinternet12y80

My strategy in these cases is usually "look for lots of facts relevant to this issue and see what stands out". The things that jump out at me from just that page:

Many American cities/states (and the entire UK in one very interesting case) have instituted or repealed gun control laws long enough ago that we can look at what happens to violent crime before and after the law is changed. In every case that they showed me, at least, places that pass gun control laws see an increase or no real change in their violent crime rate relative to national

... (read more)

6RomeoStevens12y

1st point: Regions experiencing a rise in violent crime are more likely to pass gun control laws, if the rate of rise stays approximately the same this would be evidence that gun control laws do not affect crime one way or the other. 2nd/4th point: DOJ reports approximately 20k gun deaths per year that aren't suicide. Of 8 separate studies on use of firearms by private citizens to prevent crime, the lowest number was 200k/year. This was from the study based only on police reports.

8evand12y

This also seems like a place that needs close attention to the regression fallacy. If especially high crime rate areas tend to change their gun control laws (either direction!) and then crime rates improve, that could be regression instead of cause and effect.

9Kingoftheinternet12y

That could definitely apply to a lot of the examples they presented. I'm still mystified by Washington D.C.: they already had a higher murder rate than the US average, then handguns were banned in 1975, then their murder rate tripled while the national average stayed fairly flat, then their murder rate came back down to its mid-70s level in the late 2000s, then the handgun ban was struck down. My current favored conclusion from that is "gun control laws themselves just don't matter very much, and are dwarfed by other social and cultural forces."

0NancyLebovitz12y

Even that isn't a great measure-- social changes aren't constrained by anything to rise at the same rate.

4Alsadius12y

Obvious confounding to the last fact: How many of those "someone would have died" situations would somebody actually have died in? That seems a number strongly prone to overestimation. (Of course, it's a bigger number even if you put a 90% bullshit discount on, but it's something to keep in mind)

-1Luke_A_Somers12y

1) What fraction of people are visibly armed with a gun? 2) Does that simply result in concentrating the criminals onto the other 92%? EDITED TO EXPLAIN: I misread this as committed [i]against[/i] someone visibly armed. So this was extra-confusing. Of course, I should have noticed that and gone back and been more careful.

0Kingoftheinternet12y

1) Almost zero, of course. How should that affect our interpretation of that fact? I don't understand what you mean by the second question.

-1evand12y

re: 4: I am skeptical that the fraction of reported self-defense situations in which "someone would have died" are actually situations in which someone would have died is 100%. I would ballpark it at 25%-50%, but I wouldn't be terribly shocked by any number in the range 10%-150%. Citation definitely needed on this one, especially as my "reasonable range" is wide enough to cover everything from net positive to net negative.

0Kingoftheinternet12y

They explain how they found that number here. I'm pretty impressed with their methodology, though I'm also sure you have a point about people exaggerating their chances of dying regardless of what clever study authors do.

[-]Larks12y60

In 2008, there were 14,137 homicides in the US (source), of which 9,484 involved a firearm (source). If we assume (very generously) that none of these would have happened if there were no civilian firearms, we'd save 9,484 lives.

In 2008 the value of a life was £6.9 million (source), so those deaths cost $65,400 million.

Banning guns wouldn't (supposedly) just save those though; it'd also save lives in future years. As this functions like an investment, we'll discount it by 10% (long run stock market return), and get a present value of $654,000 million.

Ther... (read more)

5prase12y

Remark about the usual policy/mind-killer issue: Your comment might provoke defensive reactions from gun control proponents who may object to your implicit identification of gun control with total ban on all guns. Since it may be good to avoid accusations of strawmanning, I'd suggest to stick to the original post's suggestion to keep one level meta: that is to say what sort of evidence is relevant rather than to present concrete pieces of evidence. That said, I miss a part analysing the relation between market value of arms and utility from having them, mostly related to value change which can happen after the ban. 1. People can be simply mistaken about their values when they buy a $5,000 gun. (Not sure how much a typical gun costs.) After the ban, those people would realise that having a gun hasn't been such a great thing and would value having a gun again much less, say $100. 2. After the ban, there is less danger from firearms and less need to have one for one's self defence. Therefore the value of arms would be lower. 3. Some subcultures are so obsessed with arms that a lot of people feel the need to own one just for signalling (that one is a patriot / not a pussy / prepared to defend one's family) even if most guns are never going to be used. After the ban the culture changes and guns are substituted by a cheaper means of signalling (knives for example). For illustration, see how absence of such considerations renders the analysis visibly faulty for a slightly different problem (all numbers in the following are made up): Abusing heroin causes death of 1,000 people in a year. Assume (very generously) that it is possible to implement an effective heroin ban which would remove all abuse. The value of life of an average person is 7 million, but it is much lower for a typical addict, say 3 million. Accounting for future deaths by 10% rate makes this 10,000 lives times 3 million which is 30 billion. Now, the present market value of all heroin in the world is 3

2Larks12y

Which is conveniently much harder to do such analysis on. Given how outrageously pro-gun-ban many of my assumptions are (e.g. I assume all the gun murders magically stop, with no substitution effect. 2nd Amendment advocates argue there would be an increase in murders!). Should I edit the comment, remove all the numbers, and prefix it with "Suppose one did this empirical work, and it came out this way. Then gun control would be bad." Would that be an improvement? Regarding your points: 1) Unless we have evidence to the contrary we should assume such errors are normally distributed with mean 0. If they had positive mean people would sell their guns for market price once they learnt their mistake, pushing the market value down to the marginal consumption value. (I suppose gun-sale regulation might make this difficult). 2) Yes, I thought of this. I think it's unlikely to be an effect in the real world, where a ban would be least-likely to disarm criminals. 3) Cheaper means of signalling? Signalling can't be cheap, or it wouldn't be signalling. I'm not sure of your point with the heroin example. This method only works if 1) your assumptions are generous in one direction and 2) the answer comes out in the other direction. (If the math had come out differently it would have been a much more conservative post, able only to identify one, prima facie stupid, argument that did not work). However, your heroin example doesn't even attempt to cover the main costs of heroin. I can't, however, see anything wrong in theory with such an approach to heroin, if someone actually did it properly. Given that all the problems you pointed out (and your accusation of bias) are on one side, and not having read anything else you've written on the subject, p=0.9 you favour an increase in gun regulation. Am I correct?

0prase12y

Depends on how exactly would you edit the comment. But if you mean removing the numbers but leaving there your conclusion that gun control would be bad, then certainly it wouldn't be an improvement. Overall I am not suggesting that you should edit the comment, once you have written it. Are you suggesting that there can't be differently priced means of signalling? (Note that I am not refering only to signalling wealth, which has to be expensive. But adherence to a specific community subculture can be signalled quite cheaply. For an extreme example consider proponents of "alternative lifestyles" who often try to visibly minimise their spending to signal their distaste for mainstream consumerism.) Not really. I don't live in the U.S. and have no opinion on that matter. In my country gun regulation exists to some extent but it is politically a non-issue. Most people (me included) don't even know how easy or difficult it is to legally buy a gun and most people (me included) don't have any opinion about gun regulation, one way or another. (Anyway, why it is interesting whether I favour increased gun regulation or not?) Regarding the heroin example: The point was to illustrate that when considering banning X, the pre-ban market value of X isn't a particularly good measure of loss resulting from the ban. The market value of existing X may greater than the loss from all negative consequences of X, but it still doesn't necessarily imply that destroying all X makes humanity worse off. If the post-ban society ceases to value X (as in the hypothetical heroin example, the former addicts no longer value the drug), it can spend resources formerly invested into X on something else. (Disclaimer: the heroin example as presented is clearly unrealistic for several reasons; I have chosen it because it is a case where it's especially clear that the good in question loses almost all its market value after an effective ban makes it unavailable.) Edit: Do you realise that this is in con

0anansi13312y

Focusing on the money makes a lot of sense to me. If we are honest with ourselves about the monetary incentives at work here, the whole discussion gets more realistic. I'm also reminded of the historical conversation having to do with the 13th amendment: In one swoop of the pen, a vast sum of money was wiped off the books, the value of all that property which was now nullified. I don't have a lot of ideas on how to make guns less profitable- unlike drugs, their high value has less to do with their legal status. But I don't think the gun lobby has got the nation's best interests in mind.

1Larks12y

The 13th amendment didn't obveously destroy any value; it just transfered value from slaveowners to slaves. Nor does this discussion have anything to do with incentives or profits. I'm just trying to quantify the value of the existing gun stock.

0drnickbone12y

An interesting economic analysis, though The Economist seems to be taking rather a different tack.

[-]fubarobfusco12y50

What would have to be true in order for increased gun control to mean fewer killings?

Why do mass murders happen? They do not happen by accident. Humans are optimizers, albeit flawed ones; we seek means to accomplish our goals. Once a human decides that killing people is a goal (terminal or instrumental), if they don't change their mind, some folks are likely to get killed. Mass murders such as those at Sandy Hook, Virginia Tech, or Columbine are premeditated — goal-directed activity, not undirected acts of chaos. The killers decided they wanted to kill, ch... (read more)

4gjm12y

I agree with JonathanLivengood and would like to state the issue a little differently: you omitted an option that we might call 2b: reduce the effectiveness of available means to kill. If guns are made less available then even if other effective means of killing are available they might be ones that lead to fewer deaths. One thing about guns of the sort used (e.g.) at Sandy Hook is that someone armed with one can kill an awful lot of people in a short period, without putting himself[1] in a vulnerable position by having to stop to reload. Note 1. Somehow, it almost always is a "himself" rather than a "herself". This paper has some interesting thoughts about why, though various stylistic features will probably be unappealing to some LW readers. [EDITED to fix a formatting glitch caused by my preferred footnote format.]

2JonathanLivengood12y

What is the evidence that 2 is out? Suppose there are five available effective means to some end. If I take away one of them, doesn't that reduce the availability of effective means to that end? Is the idea supposed to be that the various means are all so widely available that overall availability of means to the relevant end is not affected by eliminating (or greatly reducing) availability of one of them? Seems contentious to me. Moreover, what you say after the claim that 2 is out seems rather to support the claim that 2 is basically correct: poison, bombs, and knives are either practically less effective for one reason or another (hard to use, hard to practice, less destructive -- in the case of knives) or practically less available for one reason or another (poisons not widely available).

0ewbrownv12y

I think you have a point here, but there's a more fundamental problem - there doesn't seem to be much evidence that gun control affects the ability of criminals to get guns. The problem here is similar to prohibition of drugs. Guns and ammunition are widely available in many areas, are relatively easy to smuggle, and are durable goods that can be kept in operation for many decades once acquired. Also, the fact that police and other security officials need them means that they will continue to be produced and/or imported into an area with even very strict prohibition, creating many opportunities for weapons to leak out of official hands. So gun control measures are much better at disarming law-abiding citizens than criminals. Use of guns by criminals does seem to drop a bit when a nation adopts strict gun control policies for a long period of time, but the fact that the victims have been disarmed also means criminals don't need as many guns. If your goal is disarming criminals it isn't at all clear that this is a net benefit.

6fubarobfusco12y

Mass murderers such as school shooters aren't really typical criminals, though. They're very unusual criminals. Do they have access to black-market guns the way that career criminals might? Some school shooters (e.g. Wayne Lo, Seung-Hui Cho) bought their guns legally. Others (e.g. Adam Lanza) used guns belonging to family members. Harris and Klebold had another person purchase guns legally for them, but also purchased one gun illegally. Kip Kinkel was given guns by his parents, but also bought a gun that another student had stolen from a friend's father.

0Peterdjones12y

If criminals, police and piblic are all disarmed, there's just less bulletts flying around generally. There may still be plenty of crme, but there is a lot less homicide (suicide, innocent bystanders caught in corssfire, etc)

0Peterdjones12y

Two is in. Poisons and bombs arent readilly available anywhere in the first world. Knives are less effective, which is why the army and police are armed with guns. Compare this -- machete attack, 7 injured non-fatally with this--gun attack, 17 dead

[-]DanielLC12y50

my default stance is that all teachers should have concealed carry permits and mandatory shooting range time requirements

If you want to know about this specific belief, I'd suggest researching what portion of gunshot deaths occur at schools. It won't tell you whether giving teachers guns will hurt or help, but it will given you an upper limit on how much it could help. If it's only a tiny portion of the problem, don't worry about it.

8evand12y

Also: how does the time investment here compare in terms of preventable deaths to training in CPR? About other common medical problems (eg allergies)? Training in counseling for depression, suicide, abuse, rape, and pregnancy? In other words, if your goal is to prevent harm to children by better training teachers to handle it, I don't find it at all obvious that training teachers to handle an armed assailant is the best use of that (limited!) training time. I'm also not sure whether added training aimed at that is better than simply trying to train them to be better teachers (I suspect in some cases it is, but I still consider it unproven), or whether that additional training prevents enough harm to be worth the cost (again, I suspect it does, but I consider that unproven as well).

0rlpowell12y

Those are both really good points, thanks. While a school shooting is what got me thinking about this, I didn't mean to limit to that specifically; my pre-cached thought on mass murders is "if people in the crowd had been armed, the shooter would have been stopped quickly". I phrase it that way to emphasize that I have no real evidence there. Spending time training people to help those around them, in general, as a possible solution, though ... I admit that I'd honestly never thought of that. It doesn't just apply to teachers, either; one can imagine corporate "sensitivity training" that included basic lessons in how to identify/help/console a coworker who seems to be having a rough time lately. I'm not sure that we are culturally capable, even in theory, of identifying the actual impact of such a program, but it's a hell of an idea.

[-]Dr_Manhattan12y40

(my default stance is that all teachers should have concealed carry permits and mandatory shooting range time requirements)

How many kids do you think will end up being shot by teachers?
Have you ever taught in high-school?

0rlpowell12y

Heh. It was shorthand for "this is my pre-cached, non-rational response". I didn't really intend, or want, people to respond to that part, although the responses have been interesting.

[-]FiftyTwo12y30

It might be interesting to reverse the question? What benefits do you think there are from gun availability?

The ones I hear often are:

Self defence. People with guns can stop themselves being hurt. One could theoretically add up the number of times people have successfully defended themselves and compare it to gun homicides.
Pleasure gun enthusiasts get from their hobby. Seems relatively minor benefit, as they could likely get equal enjoyment from other hobbies, and most forms of regulation wouldn't affect them significantly.
Protection from government

... (read more)

6Douglas_Knight12y

The one I usually hear is deterrence. Even if guns have negative self-defense value, they may discourage certain types of attacks.

0prase12y

Can you elaborate? I am not sure if I understand what actually is the argument.

0Eugine_Nier12y

For example, I'm less likely to attempt to brake into a house if I think there's a reasonable chance of it having armed defenders.

0prase12y

But would you then say that the gun has "negative self-defense value"? That's the part by which I am confused.

0Viliam_Bur12y

I'll make a guess: It means that an average person with a gun is more likely to hurt themselves than to hurt the criminal. Yet, knowledge that given person is likely to have a gun at home, will make the criminal less likely to attack. So the gun is harmful to its owner during the act of self-defense, but increases the owner's over-all safety anyway.

0prase12y

That makes sense, thanks.

0rlpowell12y

Answering the question before reading on: I have believe in the past that more guns in the hands of ordinary, well-meaning means less violent crime, and less violent deaths, due to deterrence; "an armed society is a polite society".

0Larks12y

* Liberty as a terminal value. People being able to do what they like (subject to standard Millian proviso) is a good.

[-]kilobug12y30

My own view on gun control is that it's a kind of prisoner's dilemma equilibrium, with the "high gun" equilibrium (US) being the defect situation, and the "low gun" equilibrium (France) being the cooperate situation. And that like many cases of real life prisoner’s dilemma, an "external power" (in that case the state) enforcing the cooperation by adding an additional penalty to defection can work, but doesn't always work.

I definitely think the French situation is much saner than the US one, but I just don't know if it's realistically possible to apply it in the US.

4Eugine_Nier12y

An important point is that guns take less skill and strength to use effectively than other weapons. Thus someone without much strength or the free time to acquire the relevant skill may well prefer an environment where guns are the strategically dominant weapon. Or as the famous quote goes:

-2Peterdjones12y

If they are predominant amongst agressors and victoms alike, that obviously cancels through. In fact, if one had ones druthers, no ratioanl individual would want to be sitting at the end of an arms race.

2Eugine_Nier12y

My point is that it doesn't. Aggressors are the ones who have an advantage in combat, e.g., those who are stronger or have the free time to train. Guns reduce the "strategic inequality" between those in the biggest advantage in combat and those with the smallest.

-2Peterdjones12y

In a way that increases the chance victims will die or be injured in a confrontation. It's choosing to stand at the end of an arms race.

0Eugine_Nier12y

It also increases the chance the aggressor will die or be injured, thus reducing the motivation to become an aggressor and decreasing the chance that a confrontation occurs in the first place.

0Peterdjones12y

The UK is much more low gun than France. Gendarmes are routinely armed, bobbies are not.

-4drnickbone12y

Add a constitutional right to defect in prisoner's dilemmas, and a powerful well-funded defection lobby, and yes; it is hard to see how to get out of this. The rationalist answer (viewed from a non-US perspective) is "well just change the frigging constitution then". It's not holy writ, handed down on tablets of stone, and it is - after all - designed to be changed. There are a few other things you could usefully change while you're at it.

0Alsadius12y

That said, other less effective(but more politically palatable) options being a better way to spend your time seems a plausible case to make, at minimum. Can you imagine what it would take to get 38 states on-board to repeal the Second Amendment? Euro-style gun control in the US is about as likely as American-style gun worship in Europe. The cultures are just too different.

-6drnickbone12y

[-]CannibalSmith12y30

You may be interested in http://www.dancarlin.com/disp.php/csarchive/Show-234---Aiming-for-Effectiveness/guns-shootings-firearms

[-]OrphanWilde12y30

The first step is to isolate. As I believe culture is the most likely culprit, I'd suggest isolating by culture. Compare murder rates in the US by Japanese first and second generation immigrants to murder rates in Japan. If similar, culture is likely a substantial culprit.

1NancyLebovitz12y

If you wanted an even tighter test, check immigrants to different regions of the US, and possibly do two generations rather than just one.

[-]Suryc1112y20

Here's a report by the National Research Council of the National Academies (specifically, the Committee to Improve Research Information and Data on Firearms). It is a rigorous report that extensively discusses the issues with demonstrating a causal connection (or lack thereof)--between level of gun control and innocents killed--and generally refrains from making particular policy prescriptions.

[-]AlexanderD12y20

It will probably save time to look at recent research, which might be flawed but might also help answer some immediate questions. This might be a start; I'd quote and comment, but I broke my arm recently and typing is difficult.

[-]anansi13312y10

Since the Kern County shooting, I'm noticing that this topic (and problem) is not going away. John Stewart nailed it: http://www.thedailyshow.com/full-episodes/tue-january-8-2013-stanley-mcchrystal

-where it's a problem about framing. But it's a comedy program, not a political forum.

[disclaimer: I may be a newbie, but if I'm breaking a rule here, I can usually hear it without freaking out.]

The main problem I'm having, is that gun violence in schools makes me angry, and that makes it harder to think straight. But if gun violence in schools doesn't make som... (read more)

[-]IlyaShpitser12y10

Are we trying to be deontologist here (enough! we shouldn't have massacres of children!), or consequentialist (involuntary human deaths are bad).

If the former, this is a standard (note: I didn't say easy) causal inference problem. What you want to do, ideally, is select two reasonably large communities to serve as a test group and a control group. The communities have reasonably similar laws on the books, composition, etc. except we implement a form of gun control in one. We then check back later.

Some reasons gun control might work to stop photogenic t... (read more)

4David_Gerard12y

If you mean this one, they weren't in fact killed. As far as I know, the death toll from the Chinese attack still stands at 0. When you're a madman attacking children in a rage, a knife just isn't as effective a killing tool as an AR-15.

0rlpowell12y

You're absolutely right; this isn't the low hanging fruit for human deaths.

[-]ann07cor7y00

I always thought about this: since the dawn of violent first person shooters, our self-appointed guardians of morality have been on a crusade to warn us of the dangers of playing video games with a little edge. According to them, if you play long enough, you’ll become desensitized to graphic scenes of death and torture, emerging from behind the console as a murderous monster ready to kill his fellow humans at a mere whim. Now, as video games are getting more and more sophisticated, there’s a renewed sense of panic. But this time, it’s not from the usual su... (read more)

[-]anansi13312y00

There's a science question here, and there's an engineering question here. They are two different questions.

Science asks, what are the numbers, what are the likelihoods? And there is always going to be more study needed. Consider that people are still arguing over evolution, there may never be numbers so compelling that they convince everyone.

The engineering question asks what we could do to change things. Engineers don't get to wait for better numbers, they have to do the best they can with what they've got. We don't really know why the violent crime rat... (read more)

[-]rlpowell12y00

I'm still reeling from http://lesswrong.com/lw/g0y/gun_control_how_would_we_know/84ky?context=1#comments ; I'm noticing how in other contexts "Why are we still talking about this we have better things to do?" is obvious, but it tends not to be for me in mind-killer contexts. Unfortunately, the impact of that point on my mind is such that I'm maybe not giving this thread the attention that all of your very well-thought out answers deserve, because I've suddenly stopped caring very much.

Regardless, though, y'all certainly have lived up to my expectations as polite, reasonable, rational debaters. Well done.

[-]rlpowell12y00

Good lord. I thought I had set LW to tell me when someone replied; having not received any email I assumed this post had been ignored. 0__o

-Robin

[-]Shmi12y00

You might want to compare US and Canada, which were quite similar in terms of gun control and attitudes some 150 years ago, and then slowly diverged.

[-]Jayson_Virissimo12y100

You might want to compare US and Canada, which were quite similar in terms of gun control and attitudes some 150 years ago, and then slowly diverged.

US/Canada comparisons are extremely misleading unless you control for demographics.

0Alsadius12y

Honestly, when it comes to crime stats, I think the biggest difference is that US cities have rotted in a way that Canadian ones really haven't within the last 50 years. Canada doesn't have any Detroits or DCs to bring up our murder numbers. Also, which demographics are you controlling for here?

3Jayson_Virissimo12y

Age, ethnicity, and race are the most obvious ones.

0Alsadius12y

What's the distinction between ethnicity and race?

0Jayson_Virissimo12y

Ethnicity is a linguistic and cultural thing, while race is more of a biological thing. For obvious reasons, there is a huge overlap between the two. For some reason, the USG likes to have separate categories for Hispanic whites and non-Hispanic whites (which is why most of the datasets they release have a variable for both race and ethnicity).

0Alsadius12y

Are the way the stats tracked consistent with that distinction? For example, do they track first-generation immigrants(who are basically cultural foreigners in most cases) separately from third-generation immigrants(who are basically cultural locals in most cases) from the same racial group differently on the ethnicity tables?

-6wedrifid12y

[-]TrE12y00

Is a Causal Diagrams and Causal Models-like approach possible?

[-]buybuydandavis12y00

"[in the sole context of minimizing presumed-innocents getting randomly shot]?

What's the difference between being randomly shot, and unrandomly shot?

1TimS12y

The person pulling the trigger intends to hurt you vs. indifferent to hurting you / desires to hurt someone else and missed. Criminal law generally doesn't distinguish between these circumstances, but perhaps their causal stories are sufficiently different that some interventions would be more effective about one vs. the other. As a toy example, free marksmanship training for low-level thugs / gang-members might reduce the number of bystander deaths (and might or might not increase the number of deaths of targets of thugs).

0buybuydandavis12y

Unrandomly = intends to hurt you Randomly = indifferent to hurting you / desires to hurt someone else and missed. Is that right? You're talking about minimizing deaths of people a gunman is not trying to shoot?

1TimS12y

Actually, I think I misread rlpowell. As he seems to mean it: Unrandomly = shooter is known to victim (e.g. husband kills wife) Randomly = shooter is unknown to victim (e.g. the recent tragedy) That's a different distinction that I originally described, and a much less interesting distinction - assuming gun murders are like other murders in that most perpetrators and victims know each other. Why not? The drive-by shooting with dead target and dead innocent bystander (i.e. child down the lane hit by stray bullet) is strictly worse than the drive-by shooting with only dead target. Not that I'm aware of any worthwhile interventions to change the relative frequencies of those two events - my toy intervention is likely to have undesired knock-on consequences in the real world.

0rlpowell12y

What I was going for is the difference between wanting a particular person dead (i.e. one's wife, one's boss, etc), in which case I'd assume that access to particular weapons is irrelevant because you'll find a way, vs. wanting to kill lots of people, or to kill lots of people in a particular category (i.e. school shooting mass murders, which as I implied is how I got on this topic). It seems at least possible that weapon limitations could help limit the latter, whereas if person X really wants person Y, specifically, dead, weapon limitations seem unlikely to be relevant.

0TimS12y

As I understand it, research suggests that most desires to kill are temporary - moral philosophers might say that they aren't reflexively stable, behavioral psychologists might say that people are easily put off by trivial inconveniences. Regardless of the causal mechanism, the evidence is pretty good that unavailability of highly effective weapons prevents both random and unrandom murders. Thus, weapons limitations are likely to be relevant to all kinds of murders. Even if that isn't true, random murders are so uncommon that designing interventions specific to them is very similar to focusing medical research on curing the injuries people suffer only when struck by lighting. In short, probably a waste of attention in terms of marginal improvements.

[-]FiftyTwo12y-40

The Rationally Speaking blog on guns might provide some interesting context.

[+]Kawoomba12y-120

Moderation Log