This thread is about the 1993 Kellerman study. There’s plenty of time for all the others, including ones about suicide. The discussion about suicide in this study is only as it relates to murder-suicide.
We were talking about the chances of being killed by a gun that someone else begins into your home. I would say that felony related would qualify as breaking into an occupied home.
Then I have to wonder what you had hoped to accomplish with this thread. Yes, I note your first sentence, that you are solely interested in discussing the Kellerman study. Why that one, when there so many others (some of which I linked to) dealing with precisely the same topic? Why is the credibility of one specific study more important than the overall conclusions about the topic at hand – namely, the risks posed by firearms in the home?
Maybe I’m misreading your intent, but I’m reminded of global warming skeptics attacking Michael Mann’s paper on historical climate change reconstruction that produced the famous “hockey stick” graph. Their attacks seemed rational and analytical, much like your OP. They seemed rational because they contained elements of truth – for example, Mann’s statistical methods were indeed somewhat unorthodox. But as it turned out, using different methods produced substantially the same results, and furthermore, numerous other studies independently produced similar results. Therefore, assuming for a moment that the paper really was flawed, arguing that it was flawed would be true but totally irrelevant, and would not change the conclusion. This was my point about the Kellerman study, and introducing those other cites.
I think the fact that 84% of the case examples were without evidence of forced entry undercuts this line of supposition.
I chose this one first because it is a popular cite used by people. I’m not opposed to discussing others, but that’s not what the purpose as you note. One specific study is important to the credibility of the conclusions of that one specific study. If it is flawed, then let it be cast aside in lieu of better evidence, better data. Plus, as we can see with the current active threads in GD, there’s plenty of opportunity to discuss the meta issues, or general gun control. This isn’t that thread.
Well, “the way this study is cited” by random dudes on the internet doesn’t make the study flawed. But regardless, are you saying that it would be wrong to say “if you are worried about suicide then you shouldn’t live alone”? That because a bachelor apartment hasn’t been shown to CAUSE suicide, then that advice is misleading?
I’m not sure I follow. I am talking about the specific case where someone breaks into your home. That 84% of gun related homicides didn’t have forced entry doesn’t have any effect on the ones that did, which is the ones that I am talking about.
The paper describes the selection of control subjects, which involved starting a block from the victim’s house and interviewing people until they found a match. At least one of the neighborhoods seemed bad, but one was good. As I said, I wish the data had been broken out by counties.
I don’t know which conclusion you think can’t be drawn. That gun owning houses have higher homicide rates is clear. I don’t claim that the paper says much about guns from outside, just that there is a plausible connection between gun ownership and the introduction of outside guns.
This is really the linchpin of your argument, AFAIAC. If the test group and the control group are that dissimilar, you really don’t know how much of the difference in outcomes is caused by the thing you’re studying, and how much is caused by the underlying differences between the two groups.
Your other objection, about the firearms from outside the home, is no biggie. If the test and control groups matched up well, and the test group had 2.7 times as great a likelihood of being shot than the control group did, why would it matter where the offending weapon came from? If it came from outside, then the test group would have been 2.7 times as likely to be shot by an outside weapon. That would be one hell of a fluke.
But that doesn’t really matter, since there’s just too much daylight between test and control groups. Still, I wanted to nail down which flaw kills this study.
The upshot is, we need more, better studies. Time to ‘repeal’ the Dickey Amendment* - which actually doesn’t need repealing, it’s part of every spending bill passed by Congress since 1996. If it gets left out of the next appropriations bill (one has to be passed by March 23 - serendipitously, the day before the March For Our Lives - to avoid a shutdown), then it’s gone, at least for the time being.
*ETA: Language that mandates that “none of the funds made available for injury prevention and control at the Centers for Disease Control and Prevention (CDC) may be used to advocate or promote gun control.”
What part of the letters “If” in the beginning of my reply didn’t you understand. The first step is to show a correlation. Once you do that, you move on to investigating “why”.
So you think that the studies that showed correlation between smoking and lung cancer weren’t useful in showing that smoking increased the risk, even before the mechanism was understood?
Where did I say this was a study of causality again? Where did the authors say this?
Your example besides being stupid makes no logical sense. If tall people play basketball, and short people don’t, how could short people increase their height to become tall from playing basketball when they wouldn’t be playing to start with?
And if you cannot think of any mechanism by which a gun in the home might lead to that gun killing someone in the home, I suggest you think harder. Causation hasn’t been demonstrated, but it is plausible.
My argument was only to show a plausible connection to those who couldn’t see one. The study definitely does not give enough data to demonstrate anything.
You are right about the small number of forced entries. What is even more interesting is that over half of those forced entries were by people known to the victim.
I’d suspect many who have guns in their house for protection overestimate the probability of a stranger breaking in. I suspect that their model of using their gun is to get the bad guy - not to get their spouse or friend, which as you note is the more likely case. So not having a gun might decrease your chance of getting killed by a stranger, but increases your chance of being killed by someone you know.
That’s why the conclusion is plausible.
So, the study says guns don’t prevent crime? This argues against the one of the stated benefits of gun ownership posited by Kellerman opponents.
Yes, a study designed to demonstrate associations (correlations) can not prove causation. This is a truism. It is very difficult to prove causation, so one starts with association and moves up from there. Not unlike a police investigation.
While it does behoove people (especially researchers) to remember that correlation is not equal to causation, it is also worth remembering that causal agents ARE frequently correlated with their outcomes. Which is why we look for correlations. (This is how we know smoking causes cancer, for example. Again, despite the existence of denial by stakeholders, many lines of evidence show this.)
Many people who call out “Correlation does not equal causation!” to dispute a study they disagree with often have little problem drawing links between correlation and causation in studies whose conclusions they like (and indeed in daily life). It’s probably just human nature, but this should be guarded against in scientific endeavor.
Thanks to Voyager for excellent posts.
To V’s points about ugly sofas, I’d expect that if a reasonably designed study showed significant risks associated with ugly sofas, the first step would be attempts to replicate the data (unless Big Sofa intervened). If the data stood up, the next steps would be to attempt to tease out the issues.
Do ugly sofas drive people to suicide, or are self-loathing suicidal people likely to buy ugly sofas? Or is it homicide? Or toxic ink? Or decreased physical activity (after all, you can’t see it if you’re sitting on it)?
But, in good faith, you don’t just explain away the data without good counterdata.
They were useful in showing correlation. They did not prove causality; that came later. In terms of this gun control study, it’s now “later”. Prove the mechanism.
The authors said this -
which certainly sounds like an assumption of causality.
I’m sure there might be such mechanisms. What evidence do you have that those mechanisms exist?
There is a pretty strong correlation between being admitted to the hospital, and dying. “In light of these observations, people should be strongly discouraged from being admitted into the hospital”. Agree, or disagree? Keeping in mind that people die from iatrogenic disorders.
Regards,
Shodan
Yes, I think it would be wrong to say that if you are worried about suicides then you shouldn’t live alone. It would be a misleading presentation of the data. If the top causes of suicide are depression, poor health, and job loss, and the #50 reason is living alone, I think it would be misleading to caution against living alone as the only statement towards suicide prevention. That’s what this study does - it says that firearm ownership at home should be discouraged.
My point is that the study doesn’t bear out this supposition. What you describe may have been at play, but you can’t tell from the data. And given only a small percentage of case examples involved forced entry, drawing conclusions about break ins doesn’t stand up.
It’s not just firearms from outside the home that’s a problem, though it’s a major one. The contention is that people should be discouraged from keeping a firearm in the home based on the results of this study. Since many of the firearms used here didn’t come from the home itself, what this recommendation says is that somehow keeping a firearm in the home attracts in some way the subsequent homicide. Yes it doesn’t state a causal relationship, but it implies it quite strongly. And given that the slim majority of homicides in the case items were not committed with a firearm at all, it’s even more misleading to pin the homicide risk on the presence of a firearm at home.
I don’t think there is a single flaw that kills the study either. All the flaws combined make it so that this study shouldn’t be used in support of anything.
Were you able to suss out which of the forced entries consisted of known relationships to the victim? I didn’t see cross tab data that would show that.
I don’t think that the scenario you describe is plausible if you mean plausible to be probable or reasonable. I think as you describe is possible, but there is no indication to estimate the likelihood of that possibility.
Now this is original thinking, slipping the surly bounds of reason and boldly launching into the surrealistic syllogism. I can imagine meeting a student of formal logic, sometime in the future…
“I corresponded with Shodan, himself, back before he was so widely celebrated and admired.”
“Really? How interesting! What did you say to him?”
“Many things, lad, many. A wide range of topics…”
“And what did he say to you?”
“Mostly, he recommended the advantages of silent meditation and introspection, and never failed to assure me of his continuing admiration and regard”.
You seem to be confusing risk with causality. If the data shows that people with guns in their home are more at risk of being killed than those without, or those who smoke are more at risk of lung cancer than those who don’t, doesn’t it make sense to tell people to avoid risky behaviors even before the causality chain was known.
Say the statistics show that if you fly Wingwalker Airlines you have a 1% chance of never making it to your destination. No one knows if this is because of pilot error, bad maintenance, or their advertising on ISIS websites. Or it could be divine intervention. You still want to fly on WWA up to time the cause is found?
Funny you should mention this. Lots of patients in hospitals pick up stuff that is floating around. Hospitals are not good places to take vacations in. You really don’t want to check in with a common cold, no matter how good your insurance is.
If the risk in staying out of the hospital is greater than the risk of going in, you go in. But hospitals work hard at reducing things that might contribute to some of the risk without worrying about causality all that much. You don’t need to prove that a certain patient died from unwashed hands to force staff to wash their hands.
If you can show that the chance of getting killed from an outside attacker is greater than from a known attacker, maybe a gun is a good thing. I already said that pushers might qualify for this category. Otherwise, best not to own a gun even while the causality chain was investigated. You’d have to look at lots of police reports, though.
It was in the text
By plausible I mean we can postulate a mechanism that would make death by an outside weapon more likely with a gun in the home. I don’t see enough data to call it probable. However, what do you think is a plausible reason for the control group having fewer deaths from outside weapon than the studied group?
If I were investigating this, I’d want to look at if the friends and love interests of gun owners are more likely to own guns than the general population. If love triangles and conflict with friends are the cause of the homicide, then these same events happening with a bunch of non-gun owners would be likely to result in fewer deaths. That’s pure speculation, not a hypothesis.
Damn, I’m too old to become a sociologist.
There is also a strong correlation between going to the hospital and getting MRSA. That is not saying that you should not go to the hospital, but that you should consider the risks, make a cost benefit analysis, while at the same time, keeping up pressure on the hospitals to combat MRSA themselves through sanitation practices and patient care.
IOW, it means that you should consider if your action is more likely to cause harm than good. That doesn’t mean that in all situations, a hospital is a bad idea, just as in all situations, a gun in the home isn’t a bad idea. But, it is something that needs to be recognized that there are many situations where the hospital, or the gun, will make things worse.
One thing you certainly cannot take from this study, or any study that I’ve seen, is that a gun in the house makes you safer, you are just arguing about how much less safe having a gun makes you and your family, and whether that is worth the risk.