I Think The Gun Helps

Jonathan Haidt says that our children's brains are being rewired.

His theory, promoted most recently in his new book The Anxious Generation, is that play-based childhood has been replaced with smartphone-based childhood. Since it draws from many of the same studies, Haidt's text bears a strong resemblance to the essay from 13D Research we republished on Epsilon Theory last month. We generally agree with Haidt.

Some people disagree.

Three groups of some people, to be more precise.

There is, for example, a motley crew consisting essentially of the New Yorker, a handful of activists moonlighting as academics and a smattering of armchair Twitter sociologists who have set themselves against Haidt's position. In fairness, it isn't so much that these parties disagree with Haidt about the core assertion about childhood so much as they have decided to subject the messenger to an impromptu struggle session. You see, Haidt is a notorious Thought Criminal. He is guilty of occasionally referencing IQ. He is guilty of even considering plausible causes of the rapid rise in transgender identification among minors. He is guilty of forgetting that the Party has required us to know that human intelligence is a construct and that the rise in transgender identification can only be explained by prior oppression.

These people and their arguments are very silly. We will not revisit them again unless it is very funny that we do so.

Those who disagree with Haidt are not, however, uniformly unserious. Nature, despite converting to the much more profitable and commercially sensible left-wing editorial page business model in 2020, was once a legitimate scientific journal and probably still warrants some grudging residual respect. UC Irvine Psychology Professor Candice Odgers, who published a brief critique of Haidt's contentions in that journal, is perfectly serious.

So, too, is Marc Andreessen, whose credentials when it comes to understanding how new technologies affect society are as good as anyone's. The billionaire's name is familiar to most of you, I expect. If not, I'll simplify. Your browser-driven internet experience looks quite a bit different without his efforts at Mosaic, and the rest of your digital life looks a lot different without the billions of capital put forward by his venture capital activities over the past two decades.

The Odgers and Andreessen criticisms are not the same.

Andreessen, if I can paraphrase, sees "the phones" as the next in an illustrious line of bogeymen to which the peculiarities of a younger generation are attributed by an older generation. To the technologist, "the phones" is just "violent video games" panic all over again. He rejects Haidt's monocausal model even though Haidt's actual arguments are explicitly and exaggeratedly multicausal to a fault, but look at me trying to tell you how straw man arguments work. Andreessen then demands proof of the absence of other changing variables. It all sounds sciencey enough, although Andreessen seems happy enough to hypothesize his preferred explanatory variable - the institutionalization of oppression framing and mass-victimization by academia - with far less in the way of evidentiary support.

Odgers isn't starting from as clear a prior that a smartphone childhood couldn't possibly be responsible for declines in youth mental health, nor is she as compelled by how much evidence would be required to adjust that prior. Quite to the contrary, she is more concerned that the Right Reasons might get short shrift. That is, if we attribute too much of the rapid decline in educational attainment, mental health and executive function to smartphone-based childhood, we might miss the more important potential causes, like ... <checks notes> ... structural discrimination, economic hardship and access to guns.

That leads us to a curious, if now familiar, place. Odgers is concerned that Haidt's theory leaves too little room for leftist academic political theories, and Andreessen is concerned that Haidt's theory leaves too little room for what 20+ years of leftist academic political theories have done to a generation.

This is a feature of what we have repeatedly called the widening gyre. Haidt is "wrong" because he hasn't auto-tuned his argument about childhood to one of the two allowable polar positions on critical theory.

I am being somewhat uncharitable for effect.

Don't get me wrong. Criticizing Haidt's multicausal hypothesis by attributing declines in educational attainment and mental health among youth beginning in 2010 to structural discrimination and access to guns is objectively funny. But I'm a bit more keen on Andreessen's alternative hypothesis. Thing is, for those who have actually read Haidt's new book, that hypothesis isn't really an alternative to anything contained in the book. The rise of reverse-cognitive behavioral therapy and a generation of kids with discovered traumas and catastrophized normalcy are indispensable components of the core childhood replacement thesis. All the same, despite any hypocrisies present in the criticisms of Haidt's work and that of other researchers of these phenomena, the criticisms contain other reasonable points. We have written about many of them at length ourselves.

For example, social phenomena ARE notoriously prone to spurious correlative relationships. Studies of mental health outcomes HAVE proven hilariously difficult to replicate. Even defining "mental health" in a useful way is challenging and prone to abuse. A strong relationship, even a real relationship, doesn't obviate the potential for overdetermination. Perhaps most importantly, the presence of a demonstrable direct causal relationship rarely, if ever, implies that a particular policy response follows. No matter how much someone wants to tell you that it does. Which they will. And even if such a policy response would largely address the underlying issue, it is not always clear that the tradeoffs from implementing the policy would not dwarf the intended benefit. The path between an observation, an assertion of causality and what we should do about it is always bumpier, less linear and more complicated than the narrative package in which it is presented to us.

For all that, though, I think the right-sounding criticisms from Andreessen and Odgers are for the most part wrong. I think they do harm.

I think they are a classic case of Root Cause Paralysis.

My dad was a maintenance and reliability engineer at the largest polyethylene plants in the largest petrochemical complex in the world for much of his 42-year career.

There may not be a person on the planet who knows more about why and how furnaces and their associated equipment fail in the transformation of hydrocarbon feedstocks into little plastic pellets. He had, shall we say, a particular set of skills. His methods became a figurative and literal manual for diagnosing similar problems across his company and outside of it. They were a highly specialized application of a broader operational concept called root cause analysis.

Root cause analysis is a simple idea. Sometimes bad things happen and cause other bad things to happen. If you don't understand why the original thing happened, then your plan to fix it might just be addressing another cascading symptom of the Real Problem, or worse, something completely unrelated. You might think of it as the professional version of a three-year old who won't stop asking Why?, followed inevitably by How do you know? Unsurprisingly, this philosophy became a lesson in our home that, compounded by high-school-me being an arrogant little shit, made me a bit of a nightmare for physics and calculus teachers who occasionally wanted me to accept a formula or method without being subjected to a dozen follow-up questions.

Sorry, Mr. Willey.

In this context, root cause analysis is a method which definitionally seeks to eliminate the obfuscating influence of overdetermination wherever possible to identify a root cause against which action may be taken to eliminate a vector for harm with as little chance for misdiagnosis or unforeseen costs as possible. It's a hard problem. In a high temperature furnace, vibration, temperature, fuel irregularities, monitoring device malfunctions, design flaws and manufacturer defects can all cause similar equipment failures. Many of them can produce cascades of the other causes themselves. Many will cause failures only if the potential cascade follows a certain path. Many only under certain conditions thought by the designers to be exogenous.

None of that changes that there is no alternative to getting this right, no alternative to ruthlessly eliminating spuriously correlated variables from consideration as The Root Cause when they are simply one more symptom of the causal cascade.

This way of thinking is my patrimony. It's part of the reason I am, for example, a climate policy skeptic.

Oh, sure, I believe in anthropogenic climate change. I also believe in a massively complex climate, and I am predisposed to be more skeptical of claims that we can adequately account for the corresponding complexity, explanatory power and potential for overdetermination of such models. My concerns about the implicit (at best) incentives for publish-or-perish academic scientists to tweak models, inputs, data and generally p-hack until they can reject the null hypothesis don't make it much easier. Even so, I'd probably be comfortable with more climate policy if it weren't packaged into a narrative in which The Science frames the research findings as synonymous with the policy proposals of some congressman's draft bill. No matter how forcefully a politician or scientist tries to conflate a $250 billion, 1,500 page piece of legislation with The Science, they ain't the same thing.

My criticisms of and aversion to the flippancy of climate change policy dialogue probably sound a lot like Andreessen's criticisms of Haidt's claims on phones.

That's because they are a lot like them.

But Marc Andreessen and my dad are not the only smart men in the world. Eddie Izzard is also a smart man.

Er, sort of.

Look, because he's been very gracious about the whole thing, I think Eddie would give me grace for saying I still think of him as Eddie in the ways that make sense to me. I'll extend that grace back to him by calling him by his now-preferred Suzy. I'm with Saul from My-T-Sharp on this one. Now, Suzy isn't an expert on plastics, except inasmuch as her stage attire often included thigh-high plastic boots and drag breasts constructed of some dozen or so pounds of a siloxane polymer. Izzard is a stage comedian, and for my money probably one of the best ever to have done it. Dress to Kill is a masterpiece of high-brow low-brow observational humor, a sort of Monty-Python-meets-early-Seinfeld stand-up special. Which isn't surprising, considering that Izzard was once called "the lost Python" by John Cleese. Izzard's special and mutual Python affinity is something the troupe even played around with during the Live at Aspen reunion show back in the late 1990s.

But there's a bit from Dress to Kill that has always stuck out in my mind. Amid jokes on the Falklands, transvestite airborne battalions, the unfortunate fate of countries without flags and what the Austro-Hungarian Empire is famous for, there's a short riff on gun violence in the United States.

The National Rifle Association says that, "Guns don't kill people, people do." But I think...I think the gun helps.

Suzy Eddie Izzard, from Dress to Kill (1998)

It's a joke, of course, and a great line. But in its non-comedic form, common knowledge is that this is how dumb people think. Smart people think in root causes. Dumb people just blame what happens in the world on whatever's closest. Whatever's easiest. Monocausal explanations! The Current Thing!

I get it. I get it as an investor, a citizen and a dad. I've written at length, for example, about the importance of developing a fundamental explanation for why an investment process ought to produce alpha before we do the cargo cult hand waving of "returns analysis" and "alpha decomposition" that our industry demands of us so that we can say we are being "scientific." Every mote of my nature rages against simplistic answers to complex problems and overdetermined models. More often against the people who promote them.

But Suzy is right.

As it happens, gun violence is a devilishly useful example for bridging the logical gap between reasonable skepticism about, say, climate policy on the one hand and the role of a smartphone-based childhood in rapidly curtailing education attainment, mental health and executive function across an entire generation on the other hand.

It's also a tricky topic, absorbed as it is into the widening gyre of polarized American politics. You cannot discuss the topic of guns without being wrong in the eyes of at least half the country. Therefore, in the interest of a producing a robust mailbag segment next week, let's opt for a bit of the old spray-and-pray, instead.

Clustered tightly around one of the two permissible political points of view in these United States, there are a hundred million dyed-in-the-cruelty-free-wool liberals, progressives and leftists who are so emotionally compromised by their understandable concern about the scale of the gun violence problem in the United States that they will pretend that practically anything that might address this problem is intellectually sound. Want to listen to the Simone Biles of mental gymnastics 'explain' to you why a clause describing in plain language one reason for enumerating a human right is obviously qualifying that right? Jonesing for someone to tell you about how a massive, multi-cultural nation with 400 million guns in circulation should just do what a country with 9 million compliant citizens from a single ethnic group with limited socioeconomic stratification did? Hungry for a Very Serious Person to explain why in a country built on the rule of law it makes sense to permit unelected justices to invent arbitrary standards ("They musta meant blunderbusses and stuff!") to get around the pesky obviousness of the plain language employed in the Second Amendment?

There are a hundred million Americans positively salivating to be your Huckleberry.

Similarly huddled around the red pole are my closest political fellow-travelers, a hundred million patriots, nationalists, libertarians and the very occasional conservative (sorry, MAGA, you're about as 'conservative' as drag queen story hour at the Berkeley independent book co-op) who are so up in their feelings about fears that this horrific violence is going to end with them getting their guns taken away that they will throw their support behind practically anything that lionizes guns and minimizes their dangers. Itching for some dude with a punisher tattoo and terrible trigger discipline to seriously begin a discussion about how "good guys with guns" make us safer from criminals in public places? Fighting the inexorable urge to be informed about how knife violence in the UK proves that it's not about the gun, even though the city of Dallas had more gun murders in 2022 than all of the knife murders in England and Wales combined? Feeling nostalgic for a lecture from a baby boomer who has literally voted against funding for mental health care in at least two dozen separate elections about how gun violence isn't really so much about guns but mental health?

There are a hundred million Americans who will stop what they're doing RIGHT NOW to be your Huckleberry.

No amount of liberal word salad will change that the U.S. Constitution permits citizens to own weapons that would be useful in a military engagement against tyrants from within or without.

No amount of conservative cosplay confabs conducted in head-to-toe 5.11 gear will change that the presence of a firearm in practically any setting increases the risk of someone getting killed or maimed by a firearm.

Guns being dangerous is kinda the point.

So was letting private citizens buy and own them.

There was once a merchant captain from North Carolina who bought a 147-ton Baltimore clipper equipped with six 12-pounders and a 16-pounder long gun mounted on a swivel. He sailed up and down the Atlantic seaboard in the War of 1812, capturing British ships, maybe roughhousing some cops who got a little too nosy, and he did it all with practically zero oversight from the US government. Hell, he had their blessing - the State of New York, not so much the liberal bastion of today, was only too happy to issue him letters of marque and reprisal.

In case you were wondering, a typical 16-pounder sports a caliber of about 3.5 inches, almost 16 times the caliber of the typical AR-15 cartridge, and fires a projectile that's some 1,800 times heavier than even one of the largest you might use in the smaller firearm. Let's be generous and assume 62 grains. To be fair, the 5.56mm round exits the barrel a bit more quickly, even though the 16-pounder in this case sported a long barrel. So depending on the distance, figure that the AR-15 might manage to deliver about 0.2-0.3% of the kinetic energy of the kind of naval cannon the US government has historically been chuffed as hell for a sales bro from North Carolina to throw on his boat.

What's the point of including this little vignette? Because it's awesome, for one. Because my gun control-favoring partner and friend thought it was a bit much, for another. But mostly because, for all that, it is Team Blue that has the right of it. No, not on what the Constitution says or what federal gun laws or policies should be. Far from it. I am firmly on whatever team thinks I should get to mount a 16" gun from a retired Iowa-class battleship on the top of my barn if I want to (I want to). But they correctly spot that guns are the thing we should be talking about. Yes, even though the root cause of ANYONE non-accidentally shot with a gun in America is human behavior, both individual and collective. Yes, even though that root cause propagates in cascading effects through complex interactions with changes in social cohesion, moral values, family and community structure, policing policies and mental health resources in our country. Yes, even though we could absolutely eliminate all gun violence if we fixed all those things, if men were angels, if Santa Claus were real and tomorrow were Christmas.

Even though all of that is true, it's still the guns that we should be talking about, especially in our families, communities and schools outside the scope of sweeping national legislation.

Why? Because the presence of a gun is a self-evidently relevant variable with inherently fewer intervening variables between it and the harm caused by guns.

By comparison, when it comes to climate policy or any number of other complicated and overdetermined topics, the number of intervening and mediating variables between the operative topic and the harm is vast. The one thing you know about gun violence is that there's always a gun.

For rabid Pro-2A zealots like me, being a citizen in good faith would mean no longer playing pretend out of fear for our position. We would tell the truth. That guns are dangerous, that they make our everyday lives as Americans more dangerous, that progressives do get pretty silly and exaggerative about them, and that by actively supporting the second amendment (i.e. not just for rule of law reasons but because we think it is right and proper) we are explicitly saying that we think the threat they pose to tyrants is worth those costs.

But for the most part, we know that sounds bad and makes us look like dicks, so we don't say that. That's why this topic gives us the easiest two-part rubric for the identification of Root Cause Paralysis:

1. A self-evidently relevant variable with few intervening variables; and

1. A revealed preference on the part of the critic to use the specter of overdetermination as a tool to permanently blunt opposing arguments as yet another monocausal bogeyman put forward by people who don't understand science.

Are social media-enabled smartphones as self-evidently relevant to the declining mental health of adolescents and young adults as guns are to gun violence? Are there as few intervening variables between the proposed harm vector and the harm?

In fairness, no. But it's close. WAY closer than Andreessen or Odgers are implying, I think.

We often lean on the old Soros adage that he wasn't predicting but observing. In that spirit, let's exit the world of prediction and enter the world of observation.

We don't need an objective study to predict whether the growth in social media-connected smartphones will have a causal relationship with educational attainment. We don't even need to ask teachers for their anecdotal observations on the subject, but if you know a teacher, you really should. We can simply observe a kid staring at his phone and know that he is not looking at, listening to or understanding anything he is hearing, even if we think direct instruction in the modern school setting is sub-standard.

We don't need an objective, scientific study to predict whether the smartphone is contributing to fewer hours spent in the physical company of other non-family humans, especially leisure hours. We can simply observe that there are only 24 hours in the day (citation), and that the group-level rise in hours spent on the smartphone moved almost in lockstep with the drop in group-level hours spent in live interactions with friends.

We don't need a fully fleshed out multicausal model with exquisite explanatory power and thoroughly understood causal mechanisms to predict that a transformational socially connected technology might have transformational effects on how society functions at its most basic level. We can simply observe that in less than two decades, the entire experience of perhaps the single most powerful biological imperative - to meet a potential sexual partner - has changed completely as a demonstrable result of social media apps on smartphones. And that is to say nothing about the follow-on effects we might observe on changing social norms around casual, intentional exposure to new people at all.

Unlike the many social questions for which Root Cause Analysis and unapologetic empiricism are absolutely called for, there is a direct, unmediated transmission mechanism between social media-enabled smartphone use and all manner of observable transformative changes in society. Like the gun example, while there is a causal direction worthy of some exploration, there aren't a lot of intervening variables we need to test empirically between the question "did smartphones cause this?" and "little Susie is spending four more hours a day on her smartphone and four fewer hours playing with her friends." Likewise, I'm sure there are Root Causes in social mores and the nuclear family to explore, but "did the advent of smartphones have a lot to do with someone taking a nude of themselves on a smartphone that got shared by an ex-boyfriend via group text on a smartphone then posted on social media via a smartphone, which led to bullying in another group text on a smartphone and then in-person bullying that led to a suicide" is not sitting there waiting for some assistant professor to have GPT-4 put together just the right abstract for an exquisitely structured empirical analysis.

Now, if you want to make the argument that it has not been conclusively demonstrated that spending significantly less time with other human beings other than those with whom you share a residence is related to loneliness or mental health, um, well OK. If you want to see detailed historical explorations of how a precipitous drop in the incentive to go to public spaces with the intention of meeting new people might affect loneliness or mental health, alrighty, I guess. And if you need to see the stats before you are willing to conclude that little Billy playing Fortnite during science instruction might hurt his academic growth, more power to you. And even if you feel confident about those relationships, if you doubt that you have enough grasp on the magnitude of the social media-enabled smartphone's influence on these things relative to other factors, super.

And maybe you can look at all of these social changes that started becoming observable right about 2010-2012 when the front-facing camera emerged and Facebook's aggressive expansion of Instagram took place and decide you don't have enough evidence to conclude this isn't just correlative. Maybe you don't buy the 22 randomized control trials collated by Haidt, Twenge and Rausch which generally support mono-directional causality between social media use and depression. Maybe you think Susie avoided going out with friends and scrolled TikTok all night because; she isn't depressed because she didn't go out with friends and scrolled TikTok all night. Maybe you feel strongly that the existing arguments from Haidt and company about systematic catastrophizing and manufactured traumas from The Coddling of the American Mind seem an equally plausible cause of an overdetermined social outcome.

That's fine.

I mean it.

Speaking solely from a public policy perspective, while I respect Haidt's work and agree with it in most respects, for the time being I'm still kind of with Andreessen here. Like I wrote before, I was being uncharitable for effect. Relax, he's a big boy, he can take it. As for Haidt, any disagreement is less about evidence than it is about differences in our personal opinions about the necessary threshold for a restrictive or coercive public policy response. For example, I'm not sure I agree with Florida's new law banning social media for most minors. I think social media-empowered smartphones are immensely dangerous, but the costs of what will likely act as an effective ban or heavy restriction on information and internet access could be large, too. There ARE mediating factors between the many, many direct mechanisms triggered by social media-enabled smartphones and happiness, depression, loneliness, academic achievement and all sorts of other factors of interest to us here. It is not unreasonable on the part of Andreessen, Odgers or anyone else to want to slow our roll on public policy until we understand them better. Yes, along with any other potential causes.

But we also need to understand that no matter how destructive these technologies and the depth of their exposure within our society may be, we will and CAN never get to a place where the evidence will meet the standards being set by those infected by Root Cause Paralysis. There will always be an element of not-good-enough to any argument of concern about social media and smartphones.

Why?

Because the other major problem with narrow criticisms of the use of correlation and demands for "empirical" evidence of the linkage between harms in this case is that they presume the need to prove a linear relationship between an individual's use of smartphones relative to history or peers on the one hand, and the various hypothesized harms on the other. To the contrary, we think that many, if not most of the potential harms being done here exist not at the individual level but at the epimemetic layer of society. That is, we suspect but cannot prove that the always-on dopamine machinery of social media-connected smartphones affects the individual more through how that machinery changes social norms, conventions and the way in which information is consumed. Consider the evidence, for example, that higher social media use at the individual level is positively correlated with interactions with friends (popular people are popular, as it happens), while at the group level it appears to have a strong negative correlation.

Swearing off social media does not permit one to swear off participating in a society that has not done so. Eschewing the smartphone may reduce exposure to the dopamine machine, but it does nothing to reduce exposure to a dopamine-machine-addicted society. People will still communicate with conscientious objectors to social media in ways that they have been conditioned to by the speaking-to-a-crowd dynamic of social media, and by the conventions of non-verbal text chains.

Said another way, I think Andreessen is correct that a secular Fourth Great Awakening has taken place in academia. AND I think there is a Fourth Epimemetic Revolution taking place in communication.

The way that social media has transformed communication from one-to-one to one-to-everyone changes how we construct ideas, how we think and how we respond to memes and stories. That change is every bit as significant, I think, as the advent of spoken language, written language and the propagation of language following the invention of the printing press. We have all been rewired collectively. Again.

More to the point, many of the effects being argued by Haidt have precisely nothing to do with individual behaviors but the society created by mass adoption of these individual behaviors. Indeed, much of my hesitation about large scale response reflects the belief that the group-level nature of these effects validates some of the criticisms of public policy action. Restricting individual usage might produce less of the positive social impact (since the aggregate epimemetic effects are already baked in) but all of the negative trade-offs.

What the group-level nature of social effects doesn't validate is specious arguments like those comparing social media-enabled smartphones to fear-mongering about video game violence or prior technologies. I was a kid when Tipper and Co. were parading around the "video game violence" claims. I was in the evangelical church when the Satanic Panic was in full swing, desperate to find an underground D&D game that wouldn't get me grounded until the turn of the millennium. The idea that Haidt's claims about a social media-enabled smartphone-led childhood are comparable to these prior panics is insulting.

It is also misleading.

Most of what Haidt, his co-laborers and outside partners at organizations like Let Grow are proposing is less about relying on sweeping public policy than on collaboration with individuals, families and communities. Haidt's core contention is "Overprotection in the real world and underprotection in the virtual world" and his core prescription is that (1) parents avoid giving kids smartphones before high school, that (2) parents keep kids from social media until at least 16, that (3) parents and schools collaborate to restrict phones during the school day and that (4) parents, schools and communities work together to foster environments for more childhood independence and unsupervised play. Yes, there's one chapter in Haidt's book on how Big Tech and government may be able to help, and while there are suggestions about mandated age verification (which I, too, am not yet sold on), the vast majority of even this chapter is about creating environments for independent, unsupervised play.

These are low cost, low trade-off actions that would be entirely sensible voluntary actions at this scale regardless of whether you're on Team Fourth Great Awakening or Team Fourth Epimemetic Revolution.

Something really, really big is unequivocally happening to our kids. It's just one guy's opinion, but I have trouble imagining a single variable outside of Haidt's hypothesized multicausal hypothesis that has nearly the scale of change over this time horizon. Even if you constructed one from a cultural institution of mass victimhood, you'd be hard pressed to demonstrate a vector of its influence over children and adolescents that was not led by the social media-enabled smartphone. This bears repeating: for practically any reasonable alternative variable changing around the same time, the social media enabled smartphone is still almost certainly the primary vector for most any child or adolescent's exposure to its influence!

Maybe that's drawing the eye of dumb people like me and Suzy to a spurious correlation when we should be looking at a thousand small, interacting things producing some surprisingly synergistic effect in an overdetermined and complex social environment. Maybe I should wait for some academic to tell me that p<.05 for the relationship between ready social media access and the harms we are observing. Maybe I should be 99% confident that there are no correlated items (which coincidentally propagate exclusively through social media) to which I might assign some of the effects before I curtail my kids' exposure to social media and phones just a bit until they've learned to take risks, fail, bounce back from failure and find their identity in something other than peer approval.

Except I'm an actual parent of kids growing up right just now and doing any of that would be stupid as hell.

The standard for voluntary action on the question of smartphone-driven childhood is and should be vastly different than the standard for sweeping public policy action. The idea that it is scientific for a parent to assume NO relationship between a smartphone-based childhood and multiple correlate harms until studies have eliminated all confounders and identified the coefficient of every other variable is ludicrous.

For those with a high standard for institutional restriction on individual liberty, whether at the government or Big Tech level, I don't think that it is unreasonable that they demand more work on the costs and tradeoffs of any proposed sweeping reform. In the meantime, we do American families a disservice by actively working to dissuade parents and teachers from responsible action by straw-manning them as simplistic monocausal luddites just doing another "violent video games" panic.

Yes, the causes and explanations for what is happening to kids are overdetermined. Yes, individual-level analysis of direct relationships between the harm and social media-enabled smartphones is sometimes mixed. Yes, this is a multicausal problem. Yes, there are tradeoffs and costs to a broad public policy response that haven't yet been thoroughly examined and demonstrated.

Yes, people kill people.

But I think the gun helps.