“Figures often beguile me, particularly when I have the arranging of them myself; in which case the remark attributed to Disraeli would often apply with justice and force: ‘There are three kinds of lies: lies, damned lies, and statistics.'”
— Mark Twain
The debate over America’s response to the ongoing crisis of gun violence reached a fever pitch last month following the mass shooting in Parkland, Florida. I published an article last week calling for all sides of the issue to consider a different perspective: That in America the rates of violent crimes like homicide and robbery seem more sensitive to differences in economic inequality than to differences in the rate of gun ownership.
The article quickly generated tens of thousands of views and thousands of shares on social media. As expected it was received with a mix of praise, ambivalence, and criticism. Some critics accused me of manipulating data to arrive at my conclusion. Others questioned the validity of the source data. Some suggested that my central research question was flawed, that in choosing to study all homicides I had somehow(?) excluded gun homicides. Others wondered why I wasn’t studying mass shootings exclusively. Some disputed my choice of analysis techniques, and others went so far as to say that I have no understanding of statistics or data analysis at all — a claim that my clients, colleagues, and former professors might find amusing.
There are many people on all sides of this issue who will struggle to accept that their current opinion might be misinformed and subject to confirmation bias. They have various ways of shooting the messenger, no pun intended, and that’s fine. I’m doing this work because I’m intellectually curious about what lies at the root of the problem and am eager to revise my thinking based on what I learn, not because I expect that everyone who reads it will applaud. Some have found it helpful, and I do appreciate the many honest questions and kind words of encouragement I have received.
To achieve some scale in explaining the nuances of the work to date, I’m addressing some of the most-frequently-asked questions I’ve received. My hope is that readers will use this information to inform their own opinions on the issue and their discourse in this and other forums.
Data sources
Death and violent crime rates. The best data source that exists for the various death rates by state, intent, cause, and a multitude of other factors is the CDC’s WONDER database. The FBI also provides data on violent crimes through its Uniform Crime Reporting (UCR) database. Both of these are updated annually and provide several decades of historical data.
Firearm ownership rates. There is great variability in the reported statistics for firearm ownership in the United States:
- Estimated guns per capita. This measure, cited in Wikipedia and elsewhere, holds that there are ~101 guns per 100 people in the United States, a number far greater than that of any other country. This is a top-down estimate at the country level. No state-level data exists for this measure so it is not particularly useful in our analysis.
- Household firearm ownership rate. This measure is derived from primary research that asks respondents something along the lines of, “Do you have any guns in your house?” The measure represents the number of households out of 100 that report having guns. Several different sources for this data are widely cited:
- The journal Injury Prevention publishes an annual survey of ~4,000-5,000 US households that includes questions about gun ownership. The weakness in this survey is a sample size that is insufficient to draw any statistically significant inference at the state level. For example, a national sample of 4,000 implies only 8 or 9 respondents from Washington, DC. One person who read my work went so far as to contact the lead author of a recent study, who confirmed that the sample size was insufficient for state-level reporting; the authors discourage the study from being used in this way due to its design limitations. Nevertheless, state-level data from these studies has worked its way into popular articles on Wikipedia, Vox, and other sites.
- The CDC’s Behavioral Risk Factors Surveillance System (BRFSS) used a much larger sample size, ~200,000 households, and published household firearm ownership rates. Here is a link to 2001 BRFSS ownership data published by The Washington Post. But the CDC dropped all firearm questions from the BRFSS starting around 15 years ago, so these comprehensive state-level surveys of household firearm ownership are somewhat dated. However, the BRFSS state-level household ownership estimates are still highly correlated with the most recent Injury Prevention data. I used BRFSS data in my analysis for its larger sample size.
Central research questions
There seems to be much confusion surrounding the various measures being used to quantify the problem:
- Mass shootings are currently defined as shooting incidents having four or more deaths, excluding the perpetrator, and excluding gang or domestic violence incidents. These events occur infrequently, only 90 times in the United States between 1966 and 2012, and generally dominate the news cycle for days or weeks after each event. While the US has more mass shootings than any other country (31% of all incidents worldwide), fewer than 100 people per year die in these tragedies. The numbers are so small and the events so rare that it’s not feasible to do a meaningful study of mass shootings using quantitative statistical analysis techniques on secondary data; any well-designed study of mass shootings will likely require qualitative primary research, which is well outside my budget at present.
- Firearm homicides, including all mass shootings. There are ~10,000 of these per year in the United States.
- Non-firearm homicides committed by any means other than a firearm. There are ~7,000 of these per year.
- Homicides = firearm homicides + non-firearm homicides
- Firearm suicides, there are ~20,000 of these per year in the United States.
- Non-firearm suicides, which also total ~20,000 per year.
- Suicides, or firearm suicides + non-firearm suicides
- “Gun deaths” = firearm homicides + firearm suicides, and sometimes unintentional firearm deaths. These total ~30,000 per year, or about as many people as die in automobile accidents every year in the United States.
Here is a graphical representation of these measures, to scale, including 10-year average death rates per 100,000 people. Data is from CDC WONDER, as cited above:
Many — perhaps most — gun-control advocates focus on gun deaths, the blue area in the chart above. This is the broadest measure of the firearm-related death rate and includes gun suicides (about 2/3 of the total), gun homicides (about 1/3), and sometimes unintentional gun deaths (that little blue sliver below homicides that’s too small for a label).
Problems with the “gun deaths” measure
Studying “gun deaths” to inform policy decisions is problematic for several reasons.
First, studying how gun deaths are related to guns is about as interesting as studying how drowning is related to water. If there were no guns, there would be no gun deaths. States and countries with fewer guns generally have fewer gun deaths, and those with fewer oceans, rivers, lakes, ponds, swimming pools, bathtubs, and 5-gallon buckets generally have fewer drownings. Got it.
The problem is that factors other than guns can predict the overall suicide rate and homicide rate, and plenty of gun-substitutes exist that can be used to take one’s life or that of another.
What happens to the overall rates of homicide and suicide when there are fewer guns among a population? Do these rates go down, stay the same, or go up, and by how much?
These questions are much more interesting and valuable to study, but the gun deaths measure ignores them in how it frames the issue. Perhaps the goal of a policy change should be to reduce all homicide and suicide deaths as much as possible, not just those related to guns.
Second, there is no statistical relationship between a state’s gun homicide rate and its gun suicide rate. This observation suggests that the idea of combining these two measures into one for the purpose of informing public policy decisions may be misguided.
Here is a scatterplot of each state’s gun suicide rate vs. its gun homicide rate from the same dataset cited above:
There is no pattern here, no apparent relationship between these two variables. The factors that might influence a state’s gun suicide rate appear to be different from those that might influence its gun homicide rate. It might seem reasonable to conclude that gun homicide and gun suicide are two unrelated problems.
What about the relationship between non-gun homicide rate vs. gun homicide rate?
That’s a much more interesting correlation. A strong relationship seems to exist between these two variables. As the gun homicide rate goes up, so does the non-gun homicide rate. If you know the value of one variable for a given state, you could guess the value of the other variable within a narrow margin of error. This relationship may imply the existence of one or more common factors that explain the variance in both forms of homicide. And does this strong correlation hint at a substitution effect — meaning that if guns didn’t exist, murderers would find another way to do the deed?
Given a choice between studying gun homicides and gun suicides as a compound measure vs. all homicides as a compound measure, which do you think would be most interesting? Or the most useful, i.e. most likely to discover factors that could explain what’s going on and how we might address these problems effectively with policy change?
I think the answer is obvious. And I’m going to continue studying homicide and suicide, of all causes, as two separate and independent problems.
I think anyone talking about an easy solution to “gun deaths” is likely ignorant — however good their intentions — of just how meaningless this measure might be. It fails to appreciate the differences between its underlying problems of suicide and homicide. It fails to account for any possibility of a substitution effect in either problem.
It does make for a bigger problem to solve and easy sloganeering, supported by clean first-order statistics easily packaged into social media memes that tout the promise of a simple gun control narrative.
But managing policy to a measure that muddles the structure of its underlying problems for marketing purposes? That will lead only to disappointment.
Edit: Sam Fisher asked in a comment below for a non-gun vs. gun suicide scatterplot. Here it is, along with a matrix showing the correlation coefficients for all the measures discussed above.
Zach Mortensen 
“First, studying how gun deaths are related to guns is about as interesting as studying how drowning is related to water. If there were no guns, there would be no gun deaths. States and countries with fewer guns generally have fewer gun deaths, and those with fewer oceans, rivers, lakes, ponds, swimming pools, bathtubs, and 5-gallon buckets generally have fewer drownings. Got it.”
Your flippant dismissal of this key point here ruins your overall analysis. If indeed it is true that fewer guns = fewer gun deaths, then that’s all the proof I need that legislation will help. We can reduce the number of guns. Harder to reduce the number of rivers and oceans. And I hope you don’t use the same argument as those online that say, “should we ban cars too?” Guns, especially assault rifles, are made for a very specific purpose, to kill. It doesn’t take a bunch of shaky statistics and muddling the issues to point out, as you do, fewer guns (assault rifles), fewer deaths (mass shooting deaths).
Thanks for reading and for your comment. I think you make an excellent case for how framing has hurt our ability to have a meaningful dialogue around this issue:
“If indeed it is true that fewer guns = fewer gun deaths, then that’s all the proof I need that legislation will help.”
It’s true that fewer guns = fewer gun deaths *precisely because* that’s what the “gun deaths” metric was intended to show.
Thirty years ago the AR-15 existed, but no one had heard the term “assault rifle”. That term was designed, tested, and promoted in order to elicit the precise response you just gave: That assault rifles are doubleplus ungood, that they have no purpose in our society other than to carry out mass shootings of children and other innocent people. That if this was not the case, they wouldn’t be called “assault” rifles, would they?
But the name was created long after the thing itself, specifically to frame (read: manipulate) the way you would think about it today.
And thanks to the framing, you’re now taking the time to argue the obvious: That guns influence a measure that was designed to describe the influence of guns. That assault rifles do the thing that the name “assault rifle” was designed to say they do.
And that’s really all the proof you need?
The terms “assault rifle” and “assault weapon” are often conflated, especially by people who don’t have formal (military) firearms training. The term “assault rifle” is a military term, and in the US military refers to a rifle that, among other things, features the ability to fire in an automatic mode. This term doesn’t include the AR15s used in any of the recent shootings. The term “assault weapon” is very loosely defined. As nearly as I can tell, that term only popped into the public consciousness around the time of the debate that lead to the first “assault weapons” ban.
Very interesting analysis here, by the way.
Thanks for reading and sharing that insight!
This is not muddling the issues, this is very good, targeted statistical work that aims to provide better understanding of the factors behind crime and gun crime by treating them as two separate issues that are largely correlated. It aims to prove that if we were to somehow magically remove every gun from the US and prevent any more from coming in, it may not lead to an appreciable reduction in crime overall, which is what any legislation should be trying to prevent. Having good statistical data is imperative for crafting effective legislation that can receive bipartisan support by not disenfranchising otherwise law-abiding citizens.
Using “overall gun deaths” as a statistic is exactly the kind of muddling you describe because it treats all gun-owners as ticking timebombs that will eventually go on a mass killing spree. It is liken to the people who claim that getting rid of all black people will reduce crime by pointing at the demographics of crime statistics from Chicago or other large urban areas with large minority populations. These statistics completely ignore that this is a direct result of hundreds of years of racial discrimination and legislative oppression that has led to massive poverty, lack of education and lack of good social safety nets all of which are large factors that lead to crime. We could look at solving this by removing guns, or minorities based on “gun deaths” or the racist statistics, which could reduce rates on paper. Or we could look at tackling the underlying issues (in this case, poverty, education, safety nets) which would lead to a better quality of life for the these communities and reduced crime rates (“gun deaths” included). The author is trying to do the same thing here, to prove that we can more effectively treat crime in general with a better understanding on the underlying causes that affect both crime and gun crime, instead of celebrating when “gun deaths” are reduced, but people are still being murdered at the same rates as before.
Continuing to push legislation based on overall gun deaths, especially when a large majority of voters firmly believe in “shall not be infringed,” will lead to political suicide; the same we saw in 1994 where 50+ seats in Congress were lost to right leaning candidates after the left pushed through an assault weapons ban that has been shown to have no appreciable effect on the overall gun death rate, and frankly led to a larger amount of firearm ownership after it expired.
Lt. Kaffee – your comment about a gun’s specific purpose is to kill (presuming an negative intention, e.g. murder or suicide) can be true, cannot assume the intent behind it’s purpose. I can also claim that a gun’s specific purpose is to to kill while in the service of protecting and defending, whether it be an individual, a school, or a nation. The flaw in your argument is that if a tool can be used for ill intent, we should eliminate that tool. So let’s be sure to eliminate any tools that can be used for ill intent. Don’t like the car analogy? How about the internet? Prescription drugs? I could go on. Focusing on tools feels right, but it’s off the mark. What you’re trying to solve is ill intent. That’s hard. But that’s why we should work harder at that because reducing the motivation has no unintended negative consequences as would your premise.
Gaah – I hate autocorrect sometimes.
Fixed it for you :) Thanks for reading and for your comment, Matt.
Did you find any correlation between overall suicide rate vs firearms suicide rate? I’d be interested if the relation between the two is as strong as homicide / firearm homicide.
Thanks again for your insightful comments, Sam. I appreciate them.
I did generate a scatterplot for gun vs non-gun suicides, as well as a correlation matrix for all of these measures. I will post them. In brief: Correlation between gun suicides and gun homicides (the first scatterplot) is -0.10, between non-gun and gun homicides (second plot) is 0.83, and between non-gun and gun suicides is 0.24; so the relationship is not quite as clear when it comes to non-gun vs. gun suicides, which seems to imply that suicide is a more nuanced issue than homicide.
Interestingly, firearm ownership rate appears to be a statistically significant predictor of non-gun suicide rate in my first round of regressions, and it is an inverse relationship. Meaning that states with lower gun ownership rates have higher rates of non-gun suicides, which seems to imply that a substitution effect exists for suicide in general. Reduce the number of guns, and some people who would’ve used them for suicide will just find another way. More on that later as well.
So it turns out I couldn’t add the images to a comment; but I added them at the end of the article with a brief note explaining what they are. Thanks again for asking.
Thanks for the analysis, those are interesting results. It looks like there may be a _very_ slight correlation, but probably not enough to be statically relevant. I wonder if this can be attributed to the thought of a suicide by firearm being more quick and less painless than other methods, enough so that someone may not consider suicide if they did not have access to a firearm. If that is the case, this number may be able to be affected if guns are removed when someone has been deemed to be harmful to them self or others (similar to a psychiatric hold allowing someone to be held for treatment, or a domestic violence incident in some states). I think for a better result, one would have to look at rates pre- and post- ban to see if there was any effect.
I was curious if it’s possible that the firearm suicide rate is dis-proportionally affected by the rate of suicide from veterans. I would wager most veterans would retain access to a personal firearm after service. Not trying to write off veteran suicides as a non-issue, but I think those would also be treated differently, as it seems like many of those have glaring predications across the board (PTSD from service, lack of assistance from VA after being discharged).
The latest figures I can find here are roughly 7,500 veteran suicides / year (2016), and 66% from firearms (2014), which would put the number around 5,000. Compare that with total rates of suicide (~22,000 firearm vs ~44000 total), and it would reduce the rate from ~50% to ~46%. so not really appreciable.
Another interesting point, while looking through the NIMH data, it seems like rural states (particularly the mid and west states) have a much higher suicide rate in general, despite generally having lower total populations. I think that firearms would be much more prominent in these locations due to differing needs while living farther from more densely populated areas. I think this would skew the data pretty well when comparing rural vs urban locations, but I wouldn’t be sure how to account for any current gun restrictions that are growing more common in urban / suburban areas.
Yes, I have very similar thoughts on this. In the initial regressions the 4-factor model for robbery and homicide described in the previous post (firearm ownership rate, GDP, racial diversity, income inequality) did quite poorly at predicting the suicide rate, I believe that firearm ownership was a the only statistically significant predictor; it was very strongly positively correlated with gun suicides and negatively correlated with non-gun suicides. I want to say that the R-squares were on the order of 0.3, so something on the order of 70% of the variance in each state’s suicide rate isn’t explained by any of the 4 factors in the model.
My belief is that suicide is primarily a mental health issue, much more so than homicide or robbery. I think there is broad variance in the rates of mental illness by state and by country, with a number of underlying factors. I’m trying to find a dataset that I can use as a factor in the model. For example, the rate of prescriptions for SSRIs. I think I had that data a while back. Let me look.
This is perhaps the most disingenuous thing I have ever read. Your core problems remain, and this follow-up does nothing to erase that. I’ve addressed some of my issues with your stats – let’s look a bit more at your rhetoric.
First, from the present piece, your core argument covering both recent articles: “That in America the rates of violent crimes like homicide and robbery seem more sensitive to differences in economic inequality than to differences in the rate of gun ownership.” I do not think that it is an unfair reading to say that beyond you dismissing gun ownership as an influential factor in overall violent crime, you are advocating against any legal gun control measures beyond those that currently exist. If this is incorrect, and you are for a limited number of “statist” [your word] solutions, I would gladly be contradicted. Your arguments may have slightly more merit if you admitted the limits of your broad solution and added that, yes, of course we should look further into the data on what the impact of control measures would be and consider some options. As it stands now, though, pushing things like better background checks, bans on bump-stocks, et cetera out of the narrative as a solution to the our country’s high rate of gun violence (especially for outliers from your model – more on that in a moment) stretches your suggested impacts of reducing income inequality (100x!) beyond credibility.
“Others wondered why I wasn’t studying mass shootings exclusively.” I do not remember anyone specifically asking this, but an argument that I and others made was that you need to treat mass shootings as outliers from your model and adjust your rhetoric accordingly. Considering the number of mass shooters who are middle class, white males (relevant due to your discussions of diversity), broad plans to reduce income inequality would have no meaningful impact on these cases.
“The numbers [on mass shootings] are so small and the events so rare that it’s not feasible to do a meaningful study of mass shootings using quantitative statistical analysis techniques on secondary data; any well-designed study of mass shootings will likely require qualitative primary research, which is well outside my budget at present.” And this brings us to two huge, interrelated problems with your writing. The first is a matter of poor taste and, frankly, why I have not treated you as an intellectual or moral peer. You are using your statistical analysis to effectively gloss over the importance of children’s lives. Mass shootings represent a small number of cases, but these incidents are significant even if not statistically so. The question of what we should do in terms of gun control measures needs to include these specific cases, period, and any model that does not account for this is fundamentally flawed.
The second is that you float between saying that your work is not about mass shootings – as suggested quant / qual quote above – and trying to make it very much about mass shootings. On the latter point, quotes from the piece from ten days ago (1) and the current piece (2):
“Five years ago in the aftermath of the Sandy Hook shooting I was frustrated by the simplistic narratives being trotted out to explain why these tragedies happen.” (1)
“I published my analysis and (rather verbose) [2013, post-Sandy Hook] commentary here. The article resurfaces on Facebook, Reddit, and Twitter after every mass-shooting event, and I get a push notification about the increase in website traffic.” (1)
“Following the Pulse shooting in Orlando in July of 2016 I received an email from Jeff Dyer, one of my professors at Wharton.” (1) [This was followed by the two of you discussing gun control issues]
“The debate over America’s response to the ongoing crisis of gun violence reached a fever pitch last month following the mass shooting in Parkland, Florida.” (2)
And, of course, there is the notable fact that your pieces came out right after Sandy Hook (elementary school children as casualties) and Parkland (high schoolers), so two major mass shootings. In summary, you admit directly that you have been inspired to research and discuss gun control issues in the wake of these specific incidents sparking dialogues on how to introduce gun control measures to address mass shootings specifically. You also note your higher web traffic and sharing, so you are aware that people are sharing your data and analysis to counter arguments for gun control that are specifically in the context of the wakes of mass shootings.
[Some personal context: I originally saw your recent article because someone I know shared it on social media and used it as a launching point to berate “you f***ers” who are presently arguing for any form of gun control. My first post on your last article was originally posted as a reply to his share. The tone of that was meant more for him than you, although given some of my issues with your work, I make no apologies for transferring the tone over. I will spare you the full story about that guy accidentally discharging a firearm, negating his self-portrayal of himself as part of the “responsible gun owner” trope…]
Or, in other words, although one of the best defenses of your analysis and its methodology is that it isn’t about mass shootings, your rhetoric says otherwise. Knowing your motivation to study this at specific times and the reasons and timing of others sharing it, your repeated references to “simple” or “simplistic” gun control narratives and suggested measures is facetious. Many people – myself included – recognize mass shootings as falling outside the norm, but we find the deaths of children to be enough of a societal failing that we do not dismiss this cluster of incidents as outliers. Your snide “Control All the Guns!” image (which I am sure would disgust Allie Brosh) is you oversimplifying an issue and an unfair and untrue portrayal of people who disagree with you. I personally don’t want to “Control All the Guns,” but I do want stronger background checks along with bans on bump stocks and other modification devices at minimum. I don’t see a compelling need for center-fire, semi-automatic rifles in general, but I can get arguments for only restricting consumer purchases of those with bullet capacities over about six (and a similar ban on pistols with capacities over ten). No changes to hunting rifles, nothing stopping you from buying a pistol for self defense. Make it harder for persons who have had a mental health hold for being a danger to themselves or others, or a history of domestic abuse, to purchase a firearm. No parking lot sales, no online sales unless it is to-store delivery with a identification check on-site. I don’t buy the substitution effect argument for numerous reasons (and I would gladly back that up – I’m not now because the current response is already far too long), but even if I did, I don’t believe that things should be outright easy for mass killers, nor do I see any of the above as unreasonable impediments for responsible firearm owners.
But I digress. The point is, you admit your “tens of thousands of views and thousands of shares on social media” are being driven primarily by the mass shooting issues, so your attempts to play it both ways on whether you are writing about this matter are intellectually irresponsible. I sincerely hope that your potential clients read this and your other two pieces and make the right decision about contracting with someone who treats dead kids as a statistical anomaly and deliberately tries to steer people against solutions that could prevent or mitigate such tragedies, while simultaneously bragging about how much attention such tragedies has brought him. And your portrayal of those who disagree with you remains bunk, plain and simple.
“This is perhaps the most disingenuous thing I have ever read.”
It’s regrettable that you’ve made some significant assumptions about me from so little information, and that you’ve indulged in making them without asking a single question that might clarify your limited understanding of who I am and my detailed positions on these specific policy options you mention.
I’ve said nothing about my position on bump stocks, background checks, mental health requirements, or private party or online sales. You haven’t asked what I think about them. Yet you assert that I am “deliberately trying to steer people against solutions” and “using statistical analysis to effectively gloss over the importance of children’s lives”.
Because of these assumptions, you seem to feel justified to have “not treated [me] as an intellectual or moral peer,” and “sincerely hope” that my business fails.
Your cognitive distortions are showing, Brian. So is your hubris and your antipathy for those who you imagine disagree with you.
I say “imagine” because you and I are actually not very far apart on what we consider an appropriate legislative response. But you wouldn’t know that because you didn’t bother to ask. Why would you bother when you can just read my mind instead?
Okay, challenge accepted. What are your specific opinions about banning bump stocks and other modification kits, decreasing loopholes in background checks and improving the databases that make them work, making it difficult for persons with certain mental health issues or histories of domestic violence to obtain a firearm, across-the-board waiting periods, regulating private party sales, and banning online sales? Do you directly support any of these measures? Would you consider them?
If no, you have no real justification whatsoever for calling me out (seriously, zero), so I’m guessing yes? But I appreciate nuance, so please, details on each. And how do these fit within your analysis and narrative? How do you see these matters relating to the gun violence issue overall and the specific mass shooting issue? Why were these considerations and your support for new gun control measures not included in your prior analyses?
Thanks, it means a lot to me that you responded this way. Here’s my current thinking. As always, I’m revising my positions based on what I learn:
Ban bump stocks, etc.: Yes
The product category is an awful idea for more reasons than I have time to list. I’m skeptical about the number of lives that will be saved as there are ~300-400 total deaths caused by rifles per year and until the Vegas shooting I had never heard of a bump stock being used for anything but wasting large amounts of ammunition in the woods or desert.
Universal background checks: Yes
The background check process is simple enough even if it is technologically in the last century. Requiring the same background check for any firearm transfer (with limited exceptions noted below) makes sense and doesn’t really inconvenience anyone. I’m skeptical about the impact, though, as there seem to be a number of high-profile cases in which the killer passed a background check.
Integrated and improved federal and state databases: Yes
There remains a deep mistrust of the government having access to anything that might be used to synthesize a registry of gun owners. In this century, privacy is already dead. Accept it and move on.
Domestic violence restrictions: Yes
This should be part of the background check process. Easy win. I’d go farther and include confiscation for the duration of any order of protection, for any DV conviction until 12 months after a court-ordered abuser program has been completed and all orders of protection lifted, and a “no buy list” for these individuals during these periods.
Mental illness restrictions: Yes in principle, lots of questions on execution that may make it infeasible
This is perhaps the most difficult of all proposed policy changes because it involves protected health information, the subjective evaluation of a health care professional regarding the individual’s mental state, a registry to maintain this information and keep it updated, and a set of processes for managing who is added to the “no-buy” list, what specific diagnoses qualify people to be on the list, how we manage the people on the list, and when (and if) people are ever removed from the list. It also begs the question of whether adding someone to the list triggers confiscation of the person’s weapons or just the inability to buy new ones. A decentralized, wisdom of crowds approach would be interesting to study, one that allows the community to flag a person as a risk rather than rely exclusively on a medical professional. I love the idea of mental health restrictions in principle, but I see so many issues with execution that I wonder whether it can happen at all.
Waiting periods: Yes, with a “TSA Pre” program for frequent buyers
Perhaps a waiting period matters only for a person who is buying his or her first gun. I imagine existing gun owners could use their current weapon rather than a new one to commit most crimes, so I’m not sure what benefit a waiting period would provide. There could be a more intensive background check for a “TSA Pre” type program to skip the waiting period — my state already does this for CCW permit holders, they don’t need to go through the FBI background check because the state CCW background check is much more rigorous
Regulating private party sales: Depends on how it’s implemented
This one would be tricky to draft around the interstate commerce clause and the 10th amendment. It might need to be done indirectly with the feds creating a program and offering “incentives” for the states to pass their own laws to align with it. Regardless, there would ideally be a background check requirement and a reporting of the transaction after the fact. There might be an exemption for intra-family transfers. The carrot for reporting might be indemnity against any liability for how the transferred weapon might be used in the future; the stick might be that the owner risks being liable for any future claims that may arise. This might be used in conjunction with mandatory liability insurance, which you didn’t mention but I would also support.
Banning online sales: These are already heavily regulated in a way that is substantially equivalent to what you suggested in your last comment: I can sell you a weapon online only if any interstate transactions are run through dealers who hold a federal firearms license (FFL). If you live in Michigan and I live in Vermont, I could buy a gun from you online, but you would need to transfer it to an FFL in your state, who would charge you a fee for the transaction and ship it to another FFL in my state, who would then transfer it to me and charge a fee for the transaction. I don’t know that this system needs to change. Online sales within a single state are, I believe, considered by the ATF to be private party sales and subject only to state regulation.
I think it’s interesting to discuss opinions about these potential changes, and I think that reasonable people can find plenty of common ground about those that seem fair to all parties involved. I support most of these ideas because, to the degree that they are implemented fairly, I see little downside.
But I’m skeptical that they will actually move the needle on total homicides, because:
1. The political process tends to favor lowest-common-denominator solutions, like the assault weapons ban in 1994 that focused only on cosmetic features with the exception of banning the manufacture of new high-capacity magazines even while leaving the pre-ban mags on the market. If Congress does anything, it will likely be underwhelming by virtue of how the political process works.
2. Even if the political process can agree on a solution, and pass it, and the President signs it, we won’t know a priori where its various compromises will land and will likely have no evidence that the particular solution can be effective (see the Affordable Care Act).
3. Even if the solution can be effective, it will be challenged in court for years just like the 1994 AWB was.
4. Even if the solution holds up in court, lots of people will have to buy in, and our track record for regulating the behavior of large numbers of people is pretty poor. See prohibition, the war on drugs, immigration, ACA non-enrollment rates, etc.
5. Even if people buy in, we need to assume away substitution effects by imagining that not having access to a gun today will prevent a suicidal person from jumping off a bridge instead. We need to imagine that a domestic abuser intent on murder won’t strangle his partner instead of shooting her.
And even if all these things happen, the net effect will likely be something akin to reducing the number of guns, which I have concluded based on my most recent analysis (and recognize that you disagree) will create an effect that isn’t as powerful as that of reducing economic inequality.
So I support most of the measures discussed above because they seem fair and reasonable. And I’m skeptical that enough miracles will happen in series for these measures to move the needle.
It’s good to hear this response, and I agree on most points. However, this is still problematic when put up against your articles.
You point to the problems in passing legislation that addresses reasonable gun control measures, making these work across state lines, and avoiding the trap of legislative compromises that lead to serious complications (see: ACA). However, all of the same criticisms could be made about solutions to income inequality. There is no magic bullet solution on this front that won’t fall into the same trap – if you can prove me wrong on that point, please do.
I am all for reducing income inequality and, understanding the barriers to a quick holistic solution, I have personally done work towards decreasing the impacts of such inequality. Decreasing violence / crime is just one of the benefits that would come with broader solutions. However, I don’t see efforts to reduce income inequality and gun control measures as being either / or propositions – both can be done, and there are limits to both as solutions for reducing gun violence and violent crime in general that makes the other necessary.
The absence from the text of your support for certain, reasonable gun control measures – even if their intricacies need to be worked out – is an issue. To repeat, portraying pro-gun control discussions as simple or simplistic doesn’t hold water, especially when you see merit in certain actions. It is reductive of you to make such statements without qualifying them. The way the pieces are written now doesn’t let much daylight in even if you personally see some beyond the text, and the fact that your work is being shared widely as support for blanket anti-gun control arguments reflects this harshly.
Part of this lies in you not admitting to the limits of your study. Yeah, sure, people can share whatever they want to support their arguments, except your writing doesn’t support certain arguments. I don’t feel I need to repeat my reservations on these pieces re: mass shootings.
“And even if all these things happen, the net effect will likely be something akin to reducing the number of guns, which I have concluded based on my most recent analysis (and recognize that you disagree) will create an effect that isn’t as powerful as that of reducing economic inequality.”
Possibly, but again, differences in power here are counted in lives. Outliers that are not addressed in your primary solution are counted in lives. I feel that you are glossing over the power and efficiency of guns versus other means of violent assault, along with the impact of allowing perpetrators to carry out acts at a distance that minimizes potential harm to themselves and removes ability of victims to defend themselves (or, in some cases, their ability to just *$#%ing run and get clear of risk). The substitution effect ignores differences in magnitude between guns and other means (including how many assaults turn deadly, number of casualties), and this magnitude increases considerably when you throw high capacity and high velocity weaponry into the mix.
On the gun issue specifically – which should be the main focus when “gun” or “guns” are in the title of both articles rather than “overall violent crime” – we still have 25.2 times as many gun homicides as our 22 most economically advanced peers, and 7 times that of Canada, which is #2 among those 23 nations. Numbers at the link – tragically, when I brought this up, a news alert at the top of the page documented that shots have just been fired on a university campus. Your primary solution is not enough to erase what is a specifically American problem, and there are far more options beyond those in your texts [you allude to some briefly in the first paragraph of the last piece, only to quickly dismiss them] that can grow the net effect (deaths prevented) beyond the considerable limits of your ideas.
https://www.cnn.com/2017/10/03/americas/us-gun-statistics/index.html
Thanks for your thoughtful reply, Brian. You’ve impressed me with this.
I have many more thoughts, and links to research in other domains, that I think begin to explain why this seems to be a uniquely American problem.
I’m constrained a bit by time, and most readers seem to be able to digest no more than about 1,500 words at a time. Not a problem for you, obviously, but it does limit the pace of the discussion and constrain the format a bit. You may need to suspend disbelief for a while until I can lay out a few more of the pieces to the puzzle. But I hope you will continue to follow the conversation and keep challenging my thinking :)
Another question, and this one merits a Yes or No response:
Is it appropriate for people to share your articles when discussing societal responses to mass shootings specifically?
Add whatever justification you want after your Yes or No, but I would be grateful if you would start your reply with one of those words as a firm, unequivocal answer.
Yes, I think that any individual has the right to draw upon whatever sources he believes support his own argument.