I mean correlation doesn't mean causation. Perhaps the very reason the bought a handgun is because they thought they had some threat or were involved in activity that would increase their likely hood of getting killed.
This study looks at the broader question of people who started living with handguns in the house, from a baseline of no handgun, so many of those victims are not people purchasing weapons themselves.
The absolute numbers are low (12 killings / 100,000 pop as opposed to 8 / 100,000 in the control group).
I don't think the article claimed causation but you can't determine causation from this type of study either. There's a possibility of the data being random, as well as causality going the other way, but it is hard to temporalize getting killed causing someone to start having a handgun in the home before that point. There's also a possible third unknown and causal association that is responsible for both the killing and the prior presence of the gun. That is a definite place to do research.
If we start with those numbers and a null hyp that both the control and the "experimental" group have a mean of 10, then the chance that of finding 4 or more extra killings than in the "experimental" group is about 21%. So even assuming no p-hacking, your p-value is meaningless.
I don't think that's a valid operation. The title of the article refers to four groups of people
Group A: People who lived in a home with handguns and were killed by a gun.
Group B: People who lived in a home with handguns and were not killed by a gun.
Group C: People who lived in a home without handguns and were killed by a gun.
Group D: People who lived in a home without handguns and were not killed by a gun.
Original: "People in homes with handguns more likely to be killed (by guns)" sounds to me like "The fraction of people in homes with handguns that are killed by guns is higher than the fraction of people in homes without handguns that are killed by guns", that is "A/(A+B) > C/(C+D)".
Modified: "People who were killed (by guns) were more likely to be in homes with handguns." sounds to me like "The fraction of people who killed by guns who lived in a home with handguns is higher than the fraction of people who killed by guns who lived in a home without handguns", that is "A/(A+C) > C/(A+C)".
If, for example, you have a population with 25% of households owning guns (which describes California's population, and the study was done in California), and the risk of being killed by a gun is 12/100,000 if you live with a gun and 8/100,000 if you don't, the a sample of 1,000,000 people will have a ratio of roughly 30 : 249,970 : 60 : 749,940 in group A : B : C : D respectively.
Which means for that population, the original statement is true (people in households with guns are more likely (12/100000) to be killed by guns than people in households without guns (8/100000)) but the modified statement is false (30 of the 90 people killed by guns come from a household with guns vs 60 / 90 from a household without guns).
Edit: I suppose the reverse statement could also be
> People who were killed by guns were more likely to be in homes with handguns than people who were not killed by guns
which would be valid but to me doesn't sound like it says anything meaningfully different from
> People who lived in homes with handguns were more likely to be killed by guns than people who lived in homes without handguns.
it was a joke. the inverted statement isn't specific enough to even be wrong.
I believe it was already well-known (in a supported-by-the-data way) in the public health community that handguns at home are more risky than not having them at home, when properly integrating over all the uncertainty, and the underlying mechanism was accidental shootings unrelated to ongoing crimes.
As I read it, yes. The risk of being killed by an intruder is practically nil, while the more considerable risk of being shot by someone else in the household just requires the presence of a gun in the house.
> Those numbers suggest the risk rises 50%, but Studdert said it was actually higher: in a separate calculation designed to better account for where people live and other factors, the researchers estimated the risk was more than twice as high.
Nobody is at the mercy of news. If it bothers someone, they can stop visiting news sites, uninstall apps, discontinue their newspaper subscription or whatever. It takes more effort and restraint and some self-efficacy to do so, so it is not as convenient as occasionally complaining about the situation.
That's a good point. What might be the cause of that? Lack of self-efficacy (i.e. people not being aware that they could change the situation if they wanted to)? Lack of discipline (i.e. being aware that they could but failing to follow through)?
>If it bothers someone, they can stop visiting news sites, uninstall apps, discontinue their newspaper subscription or whatever
That only works for people not living in a society.
When you leave in a society you can suffer from societal phenomena, even if you don't participate in them.
News (not as in "things that happen and and reported" but "how the media reports news and what news it choses"), for example, can influence policy that affects you, influence your family and friends, influence your industry, and so on...
Except Ukraine, monkeypox (and covid) and the results of scientific studies are legitimately newsworthy events to cover, not "random shiny things to keep idiots occupied."
Gotta love the contrast between the authors "we only studied very specific demographics, it would be unwise to project our findings onto the whole state let alone other states" and then two paragraphs later their boss is quoted going off half cocked doing exactly that.
Only a complete idiot would think this sort of "jumping to conclusions knowing full well you don't yet have numbers to back them" behavior is anything other than harmful to a) the legitimacy of science as an abstract concept b) the reputations of the people and institutions who practice it.
The authors are trying to do the responsible thing here and their boss goes and shits on it by giving the media exactly the kind of sensational sound bytes that they will be happy to regurgitate.
Not 1:1, obviously, because guns allow killing at a distance and the general lethality of the two differs.
In Japan at least, where both guns and knives are regulated, stabbings do occur (more often than shootings, obviously) but the country still has a very low rate of homicide per capita.
Not exactly thorough. Did it cut down on the rate of successful robberies? Nonfatal assault outcomes? Animal attacks? There are more reasons than just stopping murders to have a handgun.
I would say a robbery is more successful if you can get into a house with guns. Turns out sometimes people leave the house unattended and the guns are pretty valuable.
> Did it cut down on the rate of successful robberies?
Living with a handgun owner particularly increased the risk of being shot to death in a domestic violence incident, and it did not provide any protection against being killed at home by a stranger, the researchers found.
People who lived with handgun owners “did not experience such fatal [stranger] attacks at lower rates than their neighbors in gun-free homes”, the researchers wrote, noting that stranger homicides at home were “a small minority” of the homicides observed in the study.
> Nonfatal assault outcomes? Animal attacks?
The study focused only on homicide risk and did not examine how living with a handgun owner might increase or decrease the risk of being victimized in other ways, including by nonfatal assault, home invasion, or property theft.
40 comments
[ 2.9 ms ] story [ 109 ms ] threadThe absolute numbers are low (12 killings / 100,000 pop as opposed to 8 / 100,000 in the control group).
I don't think the article claimed causation but you can't determine causation from this type of study either. There's a possibility of the data being random, as well as causality going the other way, but it is hard to temporalize getting killed causing someone to start having a handgun in the home before that point. There's also a possible third unknown and causal association that is responsible for both the killing and the prior presence of the gun. That is a definite place to do research.
R-code:
People who were killed (by guns) were more likely to be in homes with handguns.
Group A: People who lived in a home with handguns and were killed by a gun. Group B: People who lived in a home with handguns and were not killed by a gun. Group C: People who lived in a home without handguns and were killed by a gun. Group D: People who lived in a home without handguns and were not killed by a gun.
Original: "People in homes with handguns more likely to be killed (by guns)" sounds to me like "The fraction of people in homes with handguns that are killed by guns is higher than the fraction of people in homes without handguns that are killed by guns", that is "A/(A+B) > C/(C+D)".
Modified: "People who were killed (by guns) were more likely to be in homes with handguns." sounds to me like "The fraction of people who killed by guns who lived in a home with handguns is higher than the fraction of people who killed by guns who lived in a home without handguns", that is "A/(A+C) > C/(A+C)".
If, for example, you have a population with 25% of households owning guns (which describes California's population, and the study was done in California), and the risk of being killed by a gun is 12/100,000 if you live with a gun and 8/100,000 if you don't, the a sample of 1,000,000 people will have a ratio of roughly 30 : 249,970 : 60 : 749,940 in group A : B : C : D respectively.
Which means for that population, the original statement is true (people in households with guns are more likely (12/100000) to be killed by guns than people in households without guns (8/100000)) but the modified statement is false (30 of the 90 people killed by guns come from a household with guns vs 60 / 90 from a household without guns).
Edit: I suppose the reverse statement could also be
> People who were killed by guns were more likely to be in homes with handguns than people who were not killed by guns
which would be valid but to me doesn't sound like it says anything meaningfully different from
> People who lived in homes with handguns were more likely to be killed by guns than people who lived in homes without handguns.
I believe it was already well-known (in a supported-by-the-data way) in the public health community that handguns at home are more risky than not having them at home, when properly integrating over all the uncertainty, and the underlying mechanism was accidental shootings unrelated to ongoing crimes.
See also: drugs
That only works for people not living in a society.
When you leave in a society you can suffer from societal phenomena, even if you don't participate in them.
News (not as in "things that happen and and reported" but "how the media reports news and what news it choses"), for example, can influence policy that affects you, influence your family and friends, influence your industry, and so on...
This is why people don't trust the science.
The authors are trying to do the responsible thing here and their boss goes and shits on it by giving the media exactly the kind of sensational sound bytes that they will be happy to regurgitate.
In Japan at least, where both guns and knives are regulated, stabbings do occur (more often than shootings, obviously) but the country still has a very low rate of homicide per capita.
Title
> People in homes with handguns more likely to be killed
Actual Article
> people who live with handgun owners are shot to death at a higher rate than those who don’t have such weapons at home.
People with stairs in the home more likely to fall down stairs.
People with pools in the home more likely to drown.
People with guns in the home are more likely to be shot.
None of this is surprising.