Have you isolated the motivations for such obvious lying? Activism by any means? I mean, the lies are having the effect of hiding incredibly real problems, thereby making incredibly real problems harder to address. This is related to other such corporate media activist lying, but damn... why?
Yes I did. That was in a later article that I intend to port over in a few days. Look for one entitled "The Media Engine of Chaos." This problem became the main thing I wrote about over 2018 and 2019, instead of guns.
How about comparing crime rates in particular U.S. states, before and after shall-issue concealed carry laws were passed/went into effect? There DOES APPEAR to be a correlation there. Violent crime almost always decreased 5-9%. And, keeping the location the same factors out a bunch of cultural factors. This isn't "ownership" per se, but more like "practical application" of firearms, but it does seem to make us safer does it not? Am I missing something?
alaska, montana, idaho and wyoming? The only correlation that seems clear to me is that states with the most grizzly bears have the most gun ownership.
I appreciate your effort here and I think your evidence is strong. I do think you have shown that homicide rates and gun ownership doesn't have any correlation.
But saying people with guns don't kill other people, but they sure do kill themselves isn't a super strong argument for gun ownership. Can you do a suicide rate vs. gun ownership graph? It seems like there must be a correlation there.
It's not an "argument for gun ownership". It's directly speaking against the claims made by gun control groups that we had a gun homicide issue in the U.S.
The actual problem with gun deaths in the U.S. isn't the one the gun controllers complain about, and therefore it's not fixable by the gun control legislation they propose.
There was a surge of church shootings, too, until that happy day when a church elder pulled out his concealed carry and sent the perp to hell. You don't hear about them anymore.
oh that's a load of BS. It's just a rare event either way. It wasn't stopped all the sudden because of the white rhino rare GGWAG that propaganda NRA loves to sell. GG was Caucasian also, and not shot by law enforcement showing up at the scene.
As I correctly said, there was a surge of church shootings UNTIL someone in the church shot back, and there have been few if any since. You can call a demonstrable fact anything you want to call it, but that's a reflection on you, not on me or the fact.
I do consider this the crux of the issue of homicide (and we will not face it as a culture).
May I ask for your assistance to consider another facet of 'self-harm?
Can you direct me to data on the average loss in life expectancy of overdose deaths and compare/contrast it with the average loss in life expectancy caused by COVID-19 deaths?
It's clear you are no expert in statistics. 1) Any expert would be able to estimate an obvious high R2 on the plot you said 'they left out' 2) your demand of an R2 close to 1 is insane, (in other fields 0.3-0.5 is even considered a weak effect size) 3) but worst: you mention what the experts do by citing the paper, controlling for other variables but lead wtih plots looking for univariable correlations. This is INSANE. No one is arguing that gun ownership is so strong that it will wash out all other factors including violent drug violence in Columbia. The argument is that being awash in guns makes a difference. And wouldn't you know EVEN YOUR OWN DISCUSSION SHOWS THIS, as you admit that gown ownership is predictive. Now you go on to just assume that it's 'not very predictive' but that sentence does all the work of the entire article. If it is SUFFICIENTLY PREDICTIVE then your entire article falls apart. And indeed it is, and indeed it does. Let's do a causation exploration between 'being black' and 'having more guns'. Yes sending all black people away may reduce gun deaths, but it would be because you're eliminating somethign MORE FUNDAMENTAL (like gang violence or other uncontrolled variables). Now is there something MORE FUNDAMENTAL to owning a gun. Well we are controlling for these other factors that are related to gun ownership and what do you know PER SE gun ownership still makes a difference. So If it is the case that 1% increase in ownership results in 0.9% reduction in homicides then a mere reduction of 5% of guns out there saves 1350 lives EVERY YEAR. I know you don't want to believe this which is why you have to wave your hands at this point, put up a 'black people' smoke screen, because you know that this REALLY IS SIGNIFICANT and call it a 'scant correlation' and hide it halfway in your article. We don't need R2=1, we need an effect strong enough for public policy. And it turns out 0.9% can save tens or hundreds of thousands of lives given enough removal and enough time.
The main thesis of the gun control advocates is that gun ownership rates are THE causative factor in gun homicides. By your own explanation above, it's clear that it isn't the case.
Who is saying this? Who is saying this is the 'main thesis'? Who is saying that it's per se ownership and has little to do with WHO has them, HOW DANGEROUS some are HOW EASY it is to get? Can you provide citations?
You know what? Let's go with it. I'll just accept your interpretation. So we agree: I'm glad that you admit that the article really only speaks to this either strawman version OR, if you're right, some nuts from the gun control advocates that don't understand how complex the world is. Back in the real world we all know that social issues are complex and only a complete idiot (maybe the ones you allude to) would look for an 'R2 close to 1' in univariate studies. I'm sure we both agree: anyone expecting that in this complex issue is a Grade-A+ Moron.
So given the article really is against this bizarro world where no professional statisticians have never set foot, then I can see why the article might attack the idiots while simultaneously admitting (via the 0.9%) that the article is suggestive of policy proposals such as buybacks, disincentivizing purchasing, more extensive background checks among others.
Then, those arguments notwithstanding, it's completely stupid for using such horrible statistical arguments here WHICH STAND NO MATTER WHAT HE WROTE ELSEWHERE!
"No one is arguing that gun ownership is so strong that it will wash out all other factors including violent drug violence in Columbia."
Why do people make claims that are so easily disproven given a world of 8 billion people?
BTW, HWFO isn't demanding an R^2 close to 1. They're just pointing out that a perfect correlation has an R^2 of 1 AAAANNNDDDD an R^2 close to 0 means that there is no correlation. Random data.
2020 was a very unusual year, with a lot of confounders. One thing's for sure though, BLM and MAGA both need to be congratulated for avoiding any mass murders at their protests. I was sure we'd see at least one, maybe two of them, and we didn't have any. Blew me away that we could perform such a spectacle, as a country, and not end up shooting each other en masse.
If you're responding to this article while confusing the words "gun homicide" and "gun death" then you clearly didn't make it past the 2nd paragraph of the article, and there is no need for me to explain the entire article again to you in the comments. Go read the article.
Gun Deaths fall into three major categories, Homicide, Suicide and Accidents.
Call me a victim of Cognitive Bias, but about 50% of the “deaths” tied to these weapons are suicide. That’s over 50,000 per annum the last few years.
You can argue these killing devices make this a safer better world all you want. The simple fact is that a nation with the highest per capita ownership of guns will logically have a greater ratio of death from them.
Sure, a country with no guns (south korea for example) will have no gun suicides, but their overall suicide rate is something like five times greater than ours. It's almost as if suicide is its own problem with its own causes. Would you trade a x5 higher suicide rate by bridge jumping to eliminate our gun suicides? I wouldn't.
In a typical year around 2/3rds are suicide. Homicides have no bivariate correlation to gun ownership rates. States with more gun ownership will naturally have more gun suicides because a gun is a convenient way to kill yourself, but some of those suicides would have happened anyway. Anyone who considers gun suicides in their math without including the replacement factor for non-gun suicides is cheating. In fact, the replacement factor is 100% in women. That's detailed in the next article in the series.
My research (PEW and Census.GOV) shows that the states of MO, KY, OK, AR, TN, LA, MS, AL, GA, and SC have rates ranging from 20 to ≥ 25 while States in New England, NY, PA, have rates that are far below ≤10.0 to 15.0.
I also hold that the 50% of gun deaths are by suicide are largely due to easy access and availability.
However, you win because America is inundated with over 28 million legal AR15s alone, and I don’t believe we will ever see legislation or any government taking these away from owners.
I never imagined Americans living in a world of fear and paranoia, but it sure works for Fascists inspired billionaires
You are correct. You will never see AR15s taken away from anyone, because they are responsible for no more than 3% of firearm homicides. That 3% figure includes all rifles and not just the scary black ones.
Also, as Open Source Defense pointed out in this week's newsletter, 65% of all rifle sales across the entire spectrum in the past year were AR-15s. That figure even includes hunting and plinking rifles. Not only are these things ubiquitous and popular, they're so popular they've taken over the entire market.
Tell that stat to the children slaughtered with AR-15s. BTW I slept with an AR16 for 18 months in Vietnam and am quite understanding of their killing ability.
You can slaughter more kids with an ordinary police sidearm than an AR-15 because handguns are concealable. If you seriously think some school shooter is lying in bed right now thinking "Boy oh boy I sure would love to slaughter children if I had an AR-15, but all I've got is this Glock 9 so I guess I'll stay at home and play X-Box" then you need to google the word "counterfactual" and start using your brain.
What about the children slaughtered with handguns? Or knives? Or automobiles (the leading cause of death for children)? Do you not care about those children?
Sure, if as many kids or more die in auto deaths and stabbing then WTF what’s a few million AR15s and around 1,900 deaths. For you the cause or context is equivalent. So sad.
It's 40 million AR-15s and a couple dozen AR-15 child deaths. You'll find any attempt at quantifying the actual deaths from AR-15s must lead you to the mathematical conclusion that AR-15s are one of the safest things in the country.
That’s right Russ, I don’t care the least about how children die, be it in auto accidents or in a fusillade of AR15 fire in a classroom, church, or while attending a public celebration. Idiot.
“Hoplophobia” - Nope, I fully appreciate that the millions of AR15s in the hands of Americans, will never be reversed. That horse left the barn by my best guess after the Assualt Rifle Ban ended and by the time of the George Floyd murder.
Because I feel that military grade weapons have no genuine purpose in the hands of people who have no training requirement and today virtually no legal restriction, has no bearing on my feeling about guns.
1. I did a similar chart to your first one that did find a correlation between gun ownership rates and firearm homicide. Why was mine different? I posit two reasons. One is that I used data from the CDC for firearm mortality, which I would argue is more accurate since the CDC data comes from death certificates while the FBI data is only homicides that are known to law enforecment. I also do an average of multiple years, intead of one year for both gun ownership rates and firearm homicide.
2. Why wouldn't you include suicides and incidental deaths? Are they not part of the toll of gun violence? I would argue, and the data supports this, that suicide is an even bigger problem than homicide. 54% of all gun deaths are suicides. If you're trying to asses the toll gun violence is having on our society, you should probably consider the thing that makes up the majority of gun deaths.
3. Comparing the United States to places like Venezuela and Honduras is not useful. Why? Because poverty is such a strong indicator of homicide. In fact, looking at your chart comparing countries with high firearm homicide rates, I'm struck by the fact that the richest country on there (other than the US) is Panama, which has a per-capita GDP less than half that of the US. It's just below Russia, Turkey, and Malaysia according the the World Bank. The fact that the US has a comparable firearm homicide rate to these countries should tell you all you need to know about how much of an outlier the US is. When comparing the US to other countries, it's best to use high income countries, because this actually unbiases the data by controlling for things such as poverty.
It’s very useful. Venezuela and South Africa have some of the most restrictive gun laws in the planet and still have some of the highest homicide rates in the world. That seems significant. You can’t cherry pick your data to suit the narrative.
I'm not sure I follow your logic. If per-capita GDP is highly correlated to homicide and violent crime that would support controlling for that factor to isolate the effect guns are having. Plus, it's not either or. If you want to fight poverty I'm right there with you but let's also deal with the gun problem.
Why would we bother looking at guns when the highest correlating factor is GDP?
In this case, it really is either or. If GDP is the correlating factor, then we deal with that. We don't spend time and resources piddling with factors that don't correlate, such as guns.
And we certainly shouldn't automatically buy into the concept that we have a "gun problem" as an article of faith. If we have a "poverty problem", and it's the highest correlating factor, then we look at that. We don't waste our time messing around with also-ran correlates.
Appreciate your articles. I was sharing with other stats minded people, and found a study by Siegel that finds a statistically significant correlation between gun ownership rate and homicides, counteracting your piece here. I know you respect Siegel, so I was curious your thoughts. Thanks. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3828709/
Have you isolated the motivations for such obvious lying? Activism by any means? I mean, the lies are having the effect of hiding incredibly real problems, thereby making incredibly real problems harder to address. This is related to other such corporate media activist lying, but damn... why?
Yes I did. That was in a later article that I intend to port over in a few days. Look for one entitled "The Media Engine of Chaos." This problem became the main thing I wrote about over 2018 and 2019, instead of guns.
Hoplophobia.
How about comparing crime rates in particular U.S. states, before and after shall-issue concealed carry laws were passed/went into effect? There DOES APPEAR to be a correlation there. Violent crime almost always decreased 5-9%. And, keeping the location the same factors out a bunch of cultural factors. This isn't "ownership" per se, but more like "practical application" of firearms, but it does seem to make us safer does it not? Am I missing something?
The most robust analysis of that was done by Michael Siegel, and formed the backbone for this article I wrote for Open Source Defense:
https://opensourcedefense.org/blog/gun-policy-needs-a-decision-support-system
alaska, montana, idaho and wyoming? The only correlation that seems clear to me is that states with the most grizzly bears have the most gun ownership.
ban assault rifles. common sense gun reform is needed. abolish the filibuster. tear down the electoral college system.
Sometimes it's hard to tell what's parody and what isn't.
You forgot "stuff the Supreme Court"
I appreciate your effort here and I think your evidence is strong. I do think you have shown that homicide rates and gun ownership doesn't have any correlation.
But saying people with guns don't kill other people, but they sure do kill themselves isn't a super strong argument for gun ownership. Can you do a suicide rate vs. gun ownership graph? It seems like there must be a correlation there.
Yep. Give this one a go.
https://hwfo.substack.com/p/the-left-is-making-the-wrong-case?s=w
Good work, thanks.
It's not an "argument for gun ownership". It's directly speaking against the claims made by gun control groups that we had a gun homicide issue in the U.S.
The actual problem with gun deaths in the U.S. isn't the one the gun controllers complain about, and therefore it's not fixable by the gun control legislation they propose.
Yep. If you ask a wrong question, you get a wrong answer, every time.
There was a surge of church shootings, too, until that happy day when a church elder pulled out his concealed carry and sent the perp to hell. You don't hear about them anymore.
oh that's a load of BS. It's just a rare event either way. It wasn't stopped all the sudden because of the white rhino rare GGWAG that propaganda NRA loves to sell. GG was Caucasian also, and not shot by law enforcement showing up at the scene.
As I correctly said, there was a surge of church shootings UNTIL someone in the church shot back, and there have been few if any since. You can call a demonstrable fact anything you want to call it, but that's a reflection on you, not on me or the fact.
I do consider this the crux of the issue of homicide (and we will not face it as a culture).
May I ask for your assistance to consider another facet of 'self-harm?
Can you direct me to data on the average loss in life expectancy of overdose deaths and compare/contrast it with the average loss in life expectancy caused by COVID-19 deaths?
I cannot, but if you stumble into it and put together some spreadsheets on it, email them to me. I'd like to see them.
It's clear you are no expert in statistics. 1) Any expert would be able to estimate an obvious high R2 on the plot you said 'they left out' 2) your demand of an R2 close to 1 is insane, (in other fields 0.3-0.5 is even considered a weak effect size) 3) but worst: you mention what the experts do by citing the paper, controlling for other variables but lead wtih plots looking for univariable correlations. This is INSANE. No one is arguing that gun ownership is so strong that it will wash out all other factors including violent drug violence in Columbia. The argument is that being awash in guns makes a difference. And wouldn't you know EVEN YOUR OWN DISCUSSION SHOWS THIS, as you admit that gown ownership is predictive. Now you go on to just assume that it's 'not very predictive' but that sentence does all the work of the entire article. If it is SUFFICIENTLY PREDICTIVE then your entire article falls apart. And indeed it is, and indeed it does. Let's do a causation exploration between 'being black' and 'having more guns'. Yes sending all black people away may reduce gun deaths, but it would be because you're eliminating somethign MORE FUNDAMENTAL (like gang violence or other uncontrolled variables). Now is there something MORE FUNDAMENTAL to owning a gun. Well we are controlling for these other factors that are related to gun ownership and what do you know PER SE gun ownership still makes a difference. So If it is the case that 1% increase in ownership results in 0.9% reduction in homicides then a mere reduction of 5% of guns out there saves 1350 lives EVERY YEAR. I know you don't want to believe this which is why you have to wave your hands at this point, put up a 'black people' smoke screen, because you know that this REALLY IS SIGNIFICANT and call it a 'scant correlation' and hide it halfway in your article. We don't need R2=1, we need an effect strong enough for public policy. And it turns out 0.9% can save tens or hundreds of thousands of lives given enough removal and enough time.
The main thesis of the gun control advocates is that gun ownership rates are THE causative factor in gun homicides. By your own explanation above, it's clear that it isn't the case.
Who is saying this? Who is saying this is the 'main thesis'? Who is saying that it's per se ownership and has little to do with WHO has them, HOW DANGEROUS some are HOW EASY it is to get? Can you provide citations?
You know what? Let's go with it. I'll just accept your interpretation. So we agree: I'm glad that you admit that the article really only speaks to this either strawman version OR, if you're right, some nuts from the gun control advocates that don't understand how complex the world is. Back in the real world we all know that social issues are complex and only a complete idiot (maybe the ones you allude to) would look for an 'R2 close to 1' in univariate studies. I'm sure we both agree: anyone expecting that in this complex issue is a Grade-A+ Moron.
So given the article really is against this bizarro world where no professional statisticians have never set foot, then I can see why the article might attack the idiots while simultaneously admitting (via the 0.9%) that the article is suggestive of policy proposals such as buybacks, disincentivizing purchasing, more extensive background checks among others.
So everyone is in agreement.
"the article is suggestive of policy proposals such as buybacks, disincentivizing purchasing, more extensive background checks among others."
The same author has spoken clearly on the ineffectiveness of those very same policies.
No he hasn't. And where he has he's as wrong as in this very article.
He's written such on the HWFO page. I'm not sure why you insist he hasn't, and then claim where he has said it.
That statement contradicts itself.
Then, those arguments notwithstanding, it's completely stupid for using such horrible statistical arguments here WHICH STAND NO MATTER WHAT HE WROTE ELSEWHERE!
"No one is arguing that gun ownership is so strong that it will wash out all other factors including violent drug violence in Columbia."
Why do people make claims that are so easily disproven given a world of 8 billion people?
BTW, HWFO isn't demanding an R^2 close to 1. They're just pointing out that a perfect correlation has an R^2 of 1 AAAANNNDDDD an R^2 close to 0 means that there is no correlation. Random data.
Stricter federal gun laws are need in the US.
No, they aren’t.
so with 5 million new gun owners in America in 2020, the streets should be running red with blood.
2020 was a very unusual year, with a lot of confounders. One thing's for sure though, BLM and MAGA both need to be congratulated for avoiding any mass murders at their protests. I was sure we'd see at least one, maybe two of them, and we didn't have any. Blew me away that we could perform such a spectacle, as a country, and not end up shooting each other en masse.
A 2021 Pew Research article on Gun Deaths per/100,000 1921 -
https://www.pewresearch.org/short-reads/2023/04/26/what-the-data-says-about-gun-deaths-in-the-u-s/
I also found that the “Confederate” states and Bible Belt logically had higher death rates per capita then did New England and Northern states.
If you're responding to this article while confusing the words "gun homicide" and "gun death" then you clearly didn't make it past the 2nd paragraph of the article, and there is no need for me to explain the entire article again to you in the comments. Go read the article.
Gun Deaths fall into three major categories, Homicide, Suicide and Accidents.
Call me a victim of Cognitive Bias, but about 50% of the “deaths” tied to these weapons are suicide. That’s over 50,000 per annum the last few years.
You can argue these killing devices make this a safer better world all you want. The simple fact is that a nation with the highest per capita ownership of guns will logically have a greater ratio of death from them.
Sure, a country with no guns (south korea for example) will have no gun suicides, but their overall suicide rate is something like five times greater than ours. It's almost as if suicide is its own problem with its own causes. Would you trade a x5 higher suicide rate by bridge jumping to eliminate our gun suicides? I wouldn't.
In a typical year around 2/3rds are suicide. Homicides have no bivariate correlation to gun ownership rates. States with more gun ownership will naturally have more gun suicides because a gun is a convenient way to kill yourself, but some of those suicides would have happened anyway. Anyone who considers gun suicides in their math without including the replacement factor for non-gun suicides is cheating. In fact, the replacement factor is 100% in women. That's detailed in the next article in the series.
My research (PEW and Census.GOV) shows that the states of MO, KY, OK, AR, TN, LA, MS, AL, GA, and SC have rates ranging from 20 to ≥ 25 while States in New England, NY, PA, have rates that are far below ≤10.0 to 15.0.
I also hold that the 50% of gun deaths are by suicide are largely due to easy access and availability.
However, you win because America is inundated with over 28 million legal AR15s alone, and I don’t believe we will ever see legislation or any government taking these away from owners.
I never imagined Americans living in a world of fear and paranoia, but it sure works for Fascists inspired billionaires
Citation on any state having a homicide rate exceeding 20 please.
You are correct. You will never see AR15s taken away from anyone, because they are responsible for no more than 3% of firearm homicides. That 3% figure includes all rifles and not just the scary black ones.
Also, as Open Source Defense pointed out in this week's newsletter, 65% of all rifle sales across the entire spectrum in the past year were AR-15s. That figure even includes hunting and plinking rifles. Not only are these things ubiquitous and popular, they're so popular they've taken over the entire market.
Tell that stat to the children slaughtered with AR-15s. BTW I slept with an AR16 for 18 months in Vietnam and am quite understanding of their killing ability.
You can slaughter more kids with an ordinary police sidearm than an AR-15 because handguns are concealable. If you seriously think some school shooter is lying in bed right now thinking "Boy oh boy I sure would love to slaughter children if I had an AR-15, but all I've got is this Glock 9 so I guess I'll stay at home and play X-Box" then you need to google the word "counterfactual" and start using your brain.
What about the children slaughtered with handguns? Or knives? Or automobiles (the leading cause of death for children)? Do you not care about those children?
Sure, if as many kids or more die in auto deaths and stabbing then WTF what’s a few million AR15s and around 1,900 deaths. For you the cause or context is equivalent. So sad.
It's 40 million AR-15s and a couple dozen AR-15 child deaths. You'll find any attempt at quantifying the actual deaths from AR-15s must lead you to the mathematical conclusion that AR-15s are one of the safest things in the country.
https://hwfo.substack.com/p/ar-15s-are-mindbogglingly-safe
Any attempt to hang child deaths on AR-15s is an exercise in creative lying.
Your indifference to the bulk of children's deaths is palpable. You just don't care. Your hoplophobia is showing.
That’s right Russ, I don’t care the least about how children die, be it in auto accidents or in a fusillade of AR15 fire in a classroom, church, or while attending a public celebration. Idiot.
“Hoplophobia” - Nope, I fully appreciate that the millions of AR15s in the hands of Americans, will never be reversed. That horse left the barn by my best guess after the Assualt Rifle Ban ended and by the time of the George Floyd murder.
Because I feel that military grade weapons have no genuine purpose in the hands of people who have no training requirement and today virtually no legal restriction, has no bearing on my feeling about guns.
Couple of things:
1. I did a similar chart to your first one that did find a correlation between gun ownership rates and firearm homicide. Why was mine different? I posit two reasons. One is that I used data from the CDC for firearm mortality, which I would argue is more accurate since the CDC data comes from death certificates while the FBI data is only homicides that are known to law enforecment. I also do an average of multiple years, intead of one year for both gun ownership rates and firearm homicide.
Here is the chart:
https://docs.google.com/spreadsheets/d/e/2PACX-1vQa5bVwO0FP9jh4DxGNUAWsk6k8v6ZYOy8jUly8N5GG7w4IwwwPXTS8GJrgK9O7dSbAVY6rjUukbDg2/pubchart?oid=770256255&format=interactive
2. Why wouldn't you include suicides and incidental deaths? Are they not part of the toll of gun violence? I would argue, and the data supports this, that suicide is an even bigger problem than homicide. 54% of all gun deaths are suicides. If you're trying to asses the toll gun violence is having on our society, you should probably consider the thing that makes up the majority of gun deaths.
Once again, I have a chart. This is essentially an update on the Mother Jones chart with the most recent data available. It has an R^2 of 0.638: https://docs.google.com/spreadsheets/d/e/2PACX-1vRFgUUEsM_BzcWAyHXjmnG5eopBKVZun-OJ4jR2ZC2JTk_uijoehGOqz2DONRrZ4tvitrjv3AlPMNJh/pubchart?oid=707108260&format=interactive
3. Comparing the United States to places like Venezuela and Honduras is not useful. Why? Because poverty is such a strong indicator of homicide. In fact, looking at your chart comparing countries with high firearm homicide rates, I'm struck by the fact that the richest country on there (other than the US) is Panama, which has a per-capita GDP less than half that of the US. It's just below Russia, Turkey, and Malaysia according the the World Bank. The fact that the US has a comparable firearm homicide rate to these countries should tell you all you need to know about how much of an outlier the US is. When comparing the US to other countries, it's best to use high income countries, because this actually unbiases the data by controlling for things such as poverty.
It’s very useful. Venezuela and South Africa have some of the most restrictive gun laws in the planet and still have some of the highest homicide rates in the world. That seems significant. You can’t cherry pick your data to suit the narrative.
It's not cherry-picking. Proper analysis controls for variables that might color the data.
If guns are the correlating factor, then per-capita GDP would be irrelevant. In fact, it must be. But clearly your data show that it's not.
I'm not sure I follow your logic. If per-capita GDP is highly correlated to homicide and violent crime that would support controlling for that factor to isolate the effect guns are having. Plus, it's not either or. If you want to fight poverty I'm right there with you but let's also deal with the gun problem.
Why would we bother looking at guns when the highest correlating factor is GDP?
In this case, it really is either or. If GDP is the correlating factor, then we deal with that. We don't spend time and resources piddling with factors that don't correlate, such as guns.
And we certainly shouldn't automatically buy into the concept that we have a "gun problem" as an article of faith. If we have a "poverty problem", and it's the highest correlating factor, then we look at that. We don't waste our time messing around with also-ran correlates.
Appreciate your articles. I was sharing with other stats minded people, and found a study by Siegel that finds a statistically significant correlation between gun ownership rate and homicides, counteracting your piece here. I know you respect Siegel, so I was curious your thoughts. Thanks. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3828709/