What Alex Berenson’s new book gets wrong about marijuana, psychosis, and violence

Justin Sullivan/Getty Images
Finding the best ways to do good.

Alex Berenson’s Tell Your Children: The Truth About Marijuana, Mental Illness, and Violence starts early on with a personal anecdote: One night, Jacqueline Berenson, a forensic psychiatrist and Berenson’s wife, recalled a case in which a man “cut up his grandmother or set fire to his apartment.” At some point in the discussion, Jacqueline remarked, “Of course he was high, been smoking pot his whole life.”

As a journalist, this piqued Berenson’s curiosity. He claims that, previously, he saw marijuana as relatively harmless. But here was his wife, with all her expertise, linking marijuana to a gruesome crime. When he pushed back, his wife told him to look at the scientific evidence. So he did. The result is the book in which that conversation is now being retold — a book that’s gotten widespread favorable coverage in CNBC, the New Yorker, Mother Jones, and the Marshall Project, and landed op-eds from Berenson about his findings in the New York Times, the Wall Street Journal, and the Philadelphia Inquirer.

His central argument is best summarized in a few brief lines later in the book: “Marijuana causes psychosis. Psychosis causes violence. The obvious implication is that marijuana causes violence.”

I could have found this argument persuasive. I’ve become increasingly skeptical of drug legalization over the years, as I’ve reported on the opioid epidemic (caused by legal opioid painkillers), alcohol, and tobacco. I’ve written about how there are risks to marijuana that are worth taking seriously, even if one thinks that legalization is ultimately a better policy than prohibition. I’ve stopped using marijuana myself, in part because my husband had multiple experiences in which pot seemed to make his anxiety disorder flare up.

But as I read Berenson’s book, it was impossible to escape that, while a compelling read written by an experienced journalist, it is essentially an exercise in cherry-picking data and presenting correlation as causation. Observations and anecdotes, not rigorous scientific analysis, are at the core of the book’s claim that legal marijuana will cause — and, in fact, is causing — a huge rise in psychosis and violence in America.

The book largely focuses on grisly anecdotes of violent crimes committed under the influence of marijuana, the kind of “reefer madness” stories authorities and the media leaned on when they first prohibited cannabis in the 20th century.

Berenson leverages these anecdotes and limited data to argue that heavy marijuana use, spurred by the legalization of pot in several US states, is already leading to a “black tide of psychosis” and “red tide of violence.” He warns that things will only get worse as the legal pot industry grows bigger, with an incentive to stifle heavy regulations on cannabis.

In one example, he cites a recent, massive review of the evidence on marijuana’s benefits and harms from the National Academies of Sciences, Engineering, and Medicine, claiming the report, on the link between marijuana and psychosis, “declared the issue settled.”

But I read the report and wrote about it for Vox when it came out. Far from declaring this issue “settled,” the National Academies’ report was extremely careful, cautioning that marijuana’s — and marijuana addiction’s — link to psychosis “may be multidirectional and complex.” Marijuana may not cause psychosis; something else may cause both psychosis and pot use. Or the causation could go the other way: Psychotic disorders may lead to marijuana use, perhaps in an attempt to self-medicate.

“In certain societies,” the report noted, “the incidence of schizophrenia has remained stable over the past 50 years despite the introduction of cannabis into those settings.”

When discussing the report, Berenson doesn’t mention any of this. He cites only the parts of the report that are favorable to his thesis — a disservice to the massive, rigorous work of the 16-person committee of scientists, who pored over thousands of studies for the review.

And this is representative of the book as a whole.

Berenson says he named the book Tell Your Children as a tongue-in-cheek jab at his critics, because it’s the original name of the infamous movie Reefer Madness (1936), which shows people acting out violently after using marijuana. But the further I got into the book, the more it seemed like Berenson was imitating the strategy he’d meant to mock. Tell Your Children is Reefer Madness 2.0.

There are concerns about marijuana and how legalization is playing out. As the National Academies’ report makes clear, there is still a lot about cannabis that we just don’t know, including its harms and benefits. There is a risk to commercializing another product that’s addictive for some and may be harmful in other ways for others, and there may be better ways to legalize or regulate pot that minimize those risks than what we’re doing today.

But Berenson’s book, with its sensationalist claims and shoddy analysis of the evidence, doesn’t genuinely address those concerns. Tell Your Children claims to inform its readers of the “truth” about marijuana, but it instead repeatedly misleads them.

A marijuana business manager prepares for the first day of recreational sales in Denver, Colorado.
R.J. Sangosti/Denver Post via Getty Images

Berenson overstates the evidence for the link between marijuana and psychosis

Much of Berenson’s book focuses on the history of schizophrenia, psychosis, and the marijuana legalization movement. But at the crux of Berenson’s thesis is a supposed connection between marijuana, psychosis, and violence, and an argument that legalization, and the greater levels of cannabis use that will come with it, will create more violent crime in the US.

To make his case, Berenson early on reaches back to the initial wave of reefer madness, looking at the reception of marijuana in the late 19th and early 20th centuries in Mexico and India. He argues that if Mexico and India, separated by 9,000 miles, independently concluded that marijuana causes psychosis, shouldn’t we take that seriously?

He writes, “In 1901, for example, a newspaper reported on a man who attacked strangers on a street and then ‘turned on himself and with bites he tore apart his own arms until a straitjacket could be put on him … he was crazy under the influence of marihuana.’”

Mexicans “had no cultural reason to view marijuana negatively,” he later adds. “Yet they did.”

For his information on Mexico, Berenson cites Home Grown: Marijuana and the Origins of Mexico’s War on Drugs by University of Cincinnati historian Isaac Campos.

I know Campos personally. He taught me the history of the war on drugs and drug policy when I was a student at the University of Cincinnati. I read Campos’s book very early into my burgeoning interest in drug policy, and although my memory from the early 2010s is a bit fuzzy, I knew something was wrong with what Berenson wrote. I emailed Campos to confirm my hunch.

Campos cautioned that he’s only read two photocopied pages from Berenson’s book in which he’s cited. But based on that, he said that Berenson “pretty badly misrepresented” his argument. It’s not at all clear, he said, that marijuana use alone actually caused the violent outbreaks, even if that was a widely held belief in Mexico at the time.

Marijuana was mostly used in “Mexico’s most marginal environments, especially prisons and military barracks, both extremely unhospitable and violent environments,” Campos told me. And it really was widely believed at the time that marijuana caused insanity and violence, which may have created a self-fulfilling prophecy since such behavior was essentially deemed as typical while stoned. Pair all of that with the fact marijuana really can cause paranoia and anxiety (as anyone who’s used can attest to) and you get some bad stories.

But is marijuana to blame, the circumstances of its use, or a mix of these and other factors? At the very least, it’s a much more complicated story than Berenson’s one-sided portrayal.

Pictures of marijuana plants, from around the 1890s.
Universal Images Group via Getty Images

There’s a similar tale with India, where asylum reports were used to claim that marijuana was driving people to psychosis. Berenson acknowledges a government report that came out at the time, in 1894, that found the asylum reports were seriously faulty; for example, many of them ended up being cases linked to opium or alcohol, not marijuana. But Berenson dismisses this evidence, going on to focus on reports, including the faulty asylum cases, that back up his point.

He cites various studies that he suggests add scientific weight to his claims. He spends a lot of time on Sven Andréasson, a researcher from Sweden who in 1987 published in The Lancet one of the first major studies to draw a strong link between marijuana and schizophrenia.

“Based on his data and later findings, Andréasson says he believes that cannabis is responsible for between 10 percent and 15 percent of schizophrenia cases,” Berenson writes. “Few people develop schizophrenia solely because of smoking, he thinks. But many who would not have become sick do so because marijuana pushes their vulnerable brains over the edge.”

Berenson goes on along these lines, profiling researchers and studies from all over the world, from the UK to New Zealand, that purport to show a link between marijuana and psychosis. He also cites some anecdotal examples of people who recently experienced mental health issues, including Kanye West, and, without much evidence, suggests that marijuana may have been the cause of their breakdowns. (“No one seemed to connect [West’s] diagnosis to his cannabis use,” Berenson laments.)

It’s worth noting you could easily do this kind of thing in the opposite direction — looking up a bunch of studies via Google that show marijuana does not cause psychosis or related disorders. A 2018 study published in Nature Neuroscience, for example, suggested that it’s more likely that schizophrenia leads to marijuana use (potentially to cope or self-medicate), not the other way around.

But we don’t have to do this kind of cherry-picking. More rigorous reviews of the evidence have offered far more clarity than a cursory look of the evidence by a journalist — whether Berenson or me — can ever provide. That’s where the National Academies’ report comes in.

The report is very careful in its findings. It notes that there’s “substantial evidence” for an association between marijuana and psychotic disorders, and that the association is dose-dependent — greater risk correlates with heavier marijuana use. But the report also notes that the explanation for the association is unclear.

Berenson favors the idea that pot causes and worsens psychosis and psychotic disorders. The National Academies’ report, though, says that other possibilities are plausible: Maybe psychosis or psychotic disorders lead to marijuana use, or a third factor — say, genes or environment — leads to psychosis and marijuana use. It could be a mix of all these factors.

The conclusion, if there is one: “This is a complex issue, one that certainly warrants further investigation.” In other words, we don’t know yet.

Separately, the National Academies also analyzed studies on how marijuana affects symptoms of psychotic disorders. This research was more limited, although some evidence showed that a history of marijuana use may actually improve cognitive performance for people with psychotic disorders (which could explain why people with psychotic disorders self-medicate with weed, if that’s the case). But the report ultimately concluded that the evidence in this area was merely “limited” to “moderate,” so more research is necessary.

Unlike Berenson’s book, the report was conducted by more than a dozen rigorous scientists with expertise in empirical research and analysis. They looked at far more studies than Berenson cites in his book. Yet their conclusions are far tamer than his — definitely not “settled,” as he writes. That’s worth taking seriously.

This doesn’t rule out Berenson’s claims. It’s possible that he may be right to some degree, and marijuana does cause or worsen psychosis or psychotic disorders. But right now, his claims are way ahead of the evidence.

Berenson reaches with his marijuana-violence argument, falling into Reefer Madness territory

On violence, Berenson’s case is even thinner, with large swaths of the book dedicated to anecdotes of people committing violent crimes, potentially while under the influence of marijuana.

He writes, for example, about a personal trainer in Tennessee who in June 2018 killed his former boss with a hatchet. Berenson suggests that the attack was caused by marijuana, citing a previous social media post in which the attacker discussed cannabis use.

That’s … it. Even though Berenson is making an explosive claim (one that media reports and police did not make when the case was in the news, as far as I can tell), that’s all he gives us to support his suggestion that marijuana caused the attack.

This happens over and over. Berenson brings up case after case of a brutal crime, then argues that the attacker had a history of marijuana use or used cannabis shortly before the attack. There’s no evidence marijuana caused the attack; Berenson actually acknowledges this, writing that he “couldn’t always be sure cannabis played a role.” Putting the cases all together, though, he tries to make a case that there are too many of these stories for it to be a coincidence — indicating that marijuana is causing mass psychotic violence.

“Do you want more cases? Because, unfortunately, there are plenty,” he concludes. “The black tide of psychosis and the red tide of violence are rising together on a green wave, slow and steady and certain.”

A poster advertising Reefer Madness.
Hulton Archive via Getty Images

We’ve seen this before. In the late 19th century and early 20th centuries, as marijuana was introduced to new parts of the US, the media and government officials published a lot of reports exaggerating the risks of cannabis. They linked violent crimes — often involving immigrant perpetrators — to previous marijuana use. These reports culminated in Reefer Madness, the 1936 movie, and the Marihuana Tax Act of 1937, which was effectively the first federal prohibition of marijuana.

Berenson’s read of case after case feels very similar to the previous panic. There’s no consideration about how, for example, the circumstances surrounding someone’s pot use could play a role, as likely occurred with the Mexico cases. Maybe the reports are faulty, as was true with India’s asylum cases. Or maybe the attackers weren’t driven to violence by weed but were trying to cope with mental health issues that drove them to violence by using marijuana. We don’t know.

Besides, in a country where there were more than 800,000 aggravated assaults and more than 17,000 murders and non-negligent manslaughters just in 2017, and where at least 41 million people used marijuana that year, linking even dozens of cases over the years to the drug is not convincing evidence.

Berenson attempts to back up these stories with a few studies on marijuana and violence. It’s true that some individual studies have found a link, although many conflicting studies have not. This is why it’s better, when evaluating data, to rely on large, rigorous reviews of the evidence.

Such a review exists: In 2013, one was conducted by the RAND Drug Policy Research Center, commissioned by the Office of National Drug Control Policy. It concluded that “marijuana use does not induce violent crime, and the links between marijuana use and property crime are thin.”

In likely the most provocative claim in Tell Your Children, Berenson tries to argue that pot is already causing a spike in violence. Citing the same data he referenced in his New York Times op-ed, he suggests that the first four states that legalized cannabis — Alaska, Colorado, Oregon, and Washington — saw a huge rise in violent crimes afterward, which he claims was higher than a nationwide increase in violent crimes in the same time span.

Here is his data on Washington state:

In 2013, Washington had 160 murders and about 11,700 aggravated assaults, according to statewide data that the Washington Association of Sheriffs & Police Chiefs provides to the FBI for its annual national crime report. In 2017, the state had 230 murders and 13,700 aggravated assaults—an increase of about 44 percent for murders and 17 percent for aggravated assaults. That increase far outpaced the national rise in crime. Murders rose about 20 percent nationally from 2013 to 2017, and aggravated assaults about 10 percent.

Berenson doesn’t dig into the potential causes and disparities here. There isn’t a serious attempt to weed out confounding variables, with only a vague mention of some sort of “statistical analysis.” The correlation between marijuana legalization and the increase in violent crimes is given — and repeated throughout the book, as well as Berenson’s op-eds — as if the truth is obvious.

But as the old adage goes, correlation is not causation. Aaron Carroll, a pediatrician and researcher at Indiana University School of Medicine, made this point on Twitter, noting that there is a correlation between organic food sales and autism but no one seriously believes that these two are linked.

When it comes to marijuana and violence, you can pull out correlations in the opposite direction, too. As Mark Kleiman, a drug policy expert at New York University, told New York magazine, “Cannabis consumption, and especially heavy cannabis consumption, has been on the rise since 1992. Over that period, national homicide rates have fallen more than 50%.”

As with the marijuana-psychosis link, the messiness of correlation and causation is why more rigorous scientific evaluations are needed. We have that: Benjamin Hansen, an economist at the University of Oregon, put together an analysis that gauges what Colorado and Washington state’s murder rates would be if they had not legalized marijuana.

His conclusion: “While it is true that homicide rates went up in CO and WA more than they rose for the nation as a whole, the homicide rates in Colorado and Washington were actually below what the data predicted they would have been given the trends in homicides from 2000-2012,” which is almost entirely before they legalized marijuana. He added, “This suggests, at best, we can’t conclude that marijuana legalization increases violence, and perhaps even there could be small negative effects.”

Then again, maybe the lack of evidence that legalization had a big effect on violent crime shouldn’t be very surprising — based on Berenson’s admission early on in the book, where he gives some numbers for what he expects the effects of marijuana to be on psychosis and violence.

Based on work from a Dutch epidemiologist, Berenson estimates that “one extra person in 250 may develop psychosis from cannabis use” in countries with heavy use. He writes, “The United States is a big country. About 40 million Americans were born in the last decade. An increase of 0.4 percent in psychosis would mean an extra 160,000 of those kids will suffer debilitating mental illness by 2040 or so. Many thousands of those will wind up committing murder and other violent crime.”

In the context of Berenson’s book, this is meant to be really alarming — “thousands” more violent crimes is meant to sound scary. But when you consider that there are currently 330 million people in the US, and 1.2 million violent crimes in just one year nationally, those “thousands” over several decades start to seem less concerning. It’s a fraction of a fraction of a percent of all violent crime in America.

This is the worst-case scenario — one that takes Berenson’s faulty claims at face value — and it’s far from the end of the world. Intentionally or not, Berenson undermines his own claims of mass violence.

Berenson is correct that marijuana is not risk-free

Berenson is right in one respect: It is true that marijuana is not harmless.

The National Academies’ review of the evidence speaks to this. It found evidence that marijuana is linked to respiratory issues (if smoked), car crashes, lagging academic and other social achievements, and lower birth weight (if smoked during pregnancy).

There’s also the risk of addiction. As Annie Lowrey reported for the Atlantic, federal surveys suggest that one in 10 people who use marijuana will become addicted. These aren’t claims from doctors or cops. These are users self-reporting in national surveys that they struggle to stop using marijuana even as it leads them to neglect their responsibilities and has other negative impacts on their lives.

Beyond the harms on individuals, overuse and addiction are probably bad for society too. As Jon Caulkins, a drug policy expert at Carnegie Mellon University, has told me, “At some level, we know that spending more than half of your waking hours intoxicated for years and years on end is not increasing the likelihood that you’ll win a Pulitzer Prize or discover the cure for cancer.”

It remains unclear, based on federal and state data, if marijuana legalization is leading to more cases of addiction or other marijuana-related harms. But it’s at least worth keeping an eye on, given the real risks.

There is also a lot we don’t know about marijuana, including whether the far more potent strains of cannabis that people are using today will present new problems that weren’t seen before. More than anything, this is what the National Academies’ report emphasized: We have a concerning lack of data about this drug, even as we dramatically change a lot of policy around it.

On the other side, Berenson does make some good points about the overstated benefits of medical marijuana. Although advocates have claimed that marijuana treats everything from Parkinson’s disease to inflammatory bowel disease to epilepsy to PTSD, the evidence is very limited. The National Academies’ analysis found “conclusive or substantial evidence” that cannabis can help with chronic pain, chemotherapy-induced nausea and vomiting, and patient-reported multiple sclerosis spasticity symptoms, but the use of marijuana for other conditions has limited to no scientific evidence.

That’s not nothing, but it’s clear medical marijuana advocates are running way ahead of the scientific evidence with their claims. To that end, they’re making many of the same mistakes that Berenson does in his book.

A man smokes marijuana at a medical marijuana cooperative in San Francisco in 2004.
Justin Sullivan/Getty Images

There are also legitimate concerns about how the US is legalizing cannabis

Berenson is explicit that he’s writing about all of this now because of legalization. At the end of his book, he fantasizes about the marijuana movement collapsing, as it did in the 1970s when nationwide marijuana decriminalization looked increasingly certain. He writes that “in the late 1970s, once enough people experienced marijuana’s impacts up close, the tide shifted almost instantly.”

He also worries, though, that it may be too late — that legalization is already creating a huge industry with a big financial incentive to exaggerate the benefits of marijuana, downplay the risks, and fight off regulations. He claims that “full legalization will draw billions of dollars in new investment into cannabis businesses, making restrictions even more difficult.”

This wouldn’t be a new outcome in America. With the tobacco, alcohol, and opioid industries, moneyed interests have done everything to downplay the risks of their products even as tens of thousands (in the case of opioids and alcohol) or hundreds of thousands (in the case of tobacco) of Americans die each year.

But full prohibition has its own costs. Hundreds of thousands of people are arrested for marijuana each year in the US — potentially leading to a criminal record or jail time. Black Americans are far more likely to be arrested, even though they use weed at similar rates as white Americans. The black market for marijuana also effectively funds violent drug cartels and trafficking organizations around the world, allowing them to carry out their violence.

That’s why Berenson calls decriminalization, when penalties for possession are reduced but sales remain fully illegal, “a reasonable compromise.” He writes, “People shouldn’t be arrested or sent to jail for possessing marijuana. If they’re dumb enough to smoke in public, the police should take their joints and ticket them. If they’re dumb enough to be caught smoking while they’re on parole, they should be sent back to prison. But if they want to use in the privacy of their own homes, so be it.”

There are other options. A 2015 report by RAND listed a dozen alternatives to standard prohibition. Among the possibilities: legalizing possession but not sales (as Washington, DC, and Vermont have done), putting state agencies in charge of sales (as some Canadian provinces are doing, and as some states do, successfully, with alcohol), allowing only nonprofit organizations to sell pot, or permitting a handful of closely monitored for-profit companies to take part.

If we do commercial legalization, governments can also take a harder-line approach than some states are. They could impose higher taxes. They could require warning labels. They could restrict marketing. They could prohibit certain products, such as edibles that might appeal to kids and may be more likely to cause (nonfatal) overdoses. They could limit the strength of doses. They could create systems that make it harder for users to consume too much.

All of that is to say that there are real risks to marijuana, legitimate concerns about the current form of legalization America is embracing, and alternatives to what the country is doing besides the false binary choice between prohibition and commercial legalization.

But there are ways to write about all these issues while still capturing the nuance and detail they require. For that, I recommend Marijuana Legalization: What Everyone Needs to Know by Caulkins, Kleiman, and Beau Kilmer, all of whom are actual drug policy researchers and scholars. I do not recommend Berenson’s book.

Sign up for the Future Perfect newsletter. Twice a week, you’ll get a roundup of ideas and solutions for tackling our biggest challenges: improving public health, decreasing human and animal suffering, easing catastrophic risks, and — to put it simply — getting better at doing good.


Are we headed toward a “polycrisis”? The buzzword of the moment, explained.

Future Perfect

You can now sponsor refugees yourself. Here’s how.

Future Perfect

Gas stoves and the problematic politics of sacrifice

View all stories in Future Perfect
Back to top ↑