From 2001-2002 to 2012-2013, past-year cannabis use in U.S. adults more than doubled, and the share of the total population meeting DSM-IV criteria for a cannabis use disorder nearly doubled, even as the proportion of users with a disorder fell modestly.
Readers tracking how shifting laws and norms align with national patterns of cannabis use and use disorder prevalence.
30.6%of past-year cannabis users met DSM-IV criteria for a use disorder in 2012-2013, about 1 in 3.
The Backstory
Is cannabis addictive? Ask a cannabis advocate and you'll hear "less addictive than caffeine." Ask an addiction psychiatrist and you'll hear "more common than people think." Ask Deborah Hasin — the Columbia University epidemiologist who literally wrote the DSM-5 criteria for substance use disorders — and you'll get the numbers. Not opinions. Not ideology. Numbers drawn from face-to-face interviews with nearly 80,000 Americans across a decade.
Those numbers tell a story more complicated than either side of the debate wants to hear.
The Largest Substance Use Survey in America
NESARC — the National Epidemiologic Survey on Alcohol and Related Conditions — is the gold standard for understanding substance use in the United States. It is not an online poll. It is not a convenience sample of college students. It is a massive, nationally representative, face-to-face household survey conducted by trained interviewers using validated diagnostic instruments.
Hasin's study compared two waves:
- NESARC (2001-2002): 43,093 adults
- NESARC-III (2012-2013): 36,309 adults
Between those two surveys, the American cannabis landscape transformed. Medical marijuana laws expanded from 8 states to 20. Public approval of legalization shifted from minority to majority. Colorado and Washington voted to legalize recreational use. And the data showed it.
The Three Numbers That Matter
9.5%
of US adults reported using cannabis in the past year by 2012-2013 — more than double the 4.1% rate from 2001-2002. This wasn't a marginal increase. America went from roughly 1 in 25 adults using cannabis to nearly 1 in 10.
The increase was not confined to any single group. It was statistically significant across sex, age, race/ethnicity, education, marital status, income, urban/rural residence, and region. Cannabis use expanded everywhere, among everyone.
Hasin et al. (2015), JAMA Psychiatry
2.9%
of all US adults met DSM-IV criteria for cannabis use disorder in 2012-2013 — nearly double the 1.5% rate from a decade earlier. That translates to roughly 6.8 million Americans with a diagnosable cannabis use disorder.
For comparison, past-year alcohol use disorder affects approximately 13.9% of the population (NESARC-III), and past-year opioid use disorder affects roughly 0.8%. Cannabis use disorder sits between — far less common than alcohol problems, far more common than opioid addiction.
Hasin et al. (2015), JAMA Psychiatry
30.6%
of past-year cannabis users met criteria for a use disorder in 2012-2013 — roughly 1 in 3 users. This was actually a decrease from 35.6% in 2001-2002.
This is the number that both sides of the debate get wrong. The rate among users went DOWN even as the total number of people with disorders went UP. The increase in total disorders was driven entirely by more people using cannabis, not by cannabis becoming more dangerous per user. New, more casual users were entering the population and diluting the disorder rate — but the sheer number of additional users meant more total cases.
Hasin et al. (2015), JAMA Psychiatry
The Nuance Everyone Misses
Cannabis Advocates
Focus on the decreased rate among users (35.6% → 30.6%) as evidence that cannabis is becoming safer
The absolute number of Americans with cannabis use disorder nearly doubled. A lower rate applied to a much larger population still means millions more people with a diagnosable problem.
The per-user risk did decline. More casual users entering the pool brought the average rate down.
Anti-Cannabis Groups
Focus on the doubled population prevalence (1.5% → 2.9%) as evidence that legalization causes addiction
The per-user risk actually decreased. The data doesn't show cannabis becoming more addictive — it shows more people using it. Most of the study period (2001-2013) predates any recreational legalization.
The total number of CUD cases did increase dramatically across nearly all demographic groups.
The data shows exactly what public health economists would predict: when access to a substance increases, total consumption increases, and total harms increase — even if the risk per user stays flat or declines. This is not unique to cannabis. It is the basic epidemiology of any widely used substance.
Hasin et al. (2015); Grucza et al. (2016) commentary; Hasin (2016) reply
How Addictive Is Cannabis, Really?
Hasin's data fits into a larger comparative framework. The best reference point remains Anthony et al. (1994), which used the earlier National Comorbidity Survey to estimate lifetime dependence risk among people who had ever tried each substance:
The Controversy: Are the Numbers Real?
Shortly after publication, Richard Grucza and colleagues raised a pointed methodological question in a JAMA Psychiatry commentary: were the increases real, or were people simply more willing to admit cannabis use in 2012-2013 due to changing social norms?
The concern was legitimate. The two NESARC waves used different interviewing organizations — Wave 1 used US Census Bureau employees, Wave 3 used a private contractor. It is plausible that people would be more forthcoming about illegal drug use with a private interviewer than with a government employee.
How They Did It
How Hasin Validated the Findings
Grucza et al. noted that changing social acceptability and different interview methodology could inflate apparent increases. If people simply disclosed more in 2012-2013, the true prevalence change might be smaller.
Hasin pointed to VA medical records showing a 50%+ increase in cannabis use disorder diagnoses from 2002-2009. These are clinician-assessed diagnoses in a healthcare system, not self-report surveys — and they showed the same upward trend.
SAMHSA's Drug Abuse Warning Network showed a 62% increase in marijuana involvement in emergency department visits from 2004-2011. ER visits are objective events, not subject to social desirability bias.
The Monitoring the Future survey, the National Survey on Drug Use and Health, and state-level prevalence data all showed increases in the same direction and roughly similar magnitude. It would be remarkable if every data source happened to produce the same bias.
Hasin (2016), JAMA Psychiatry reply; Grucza et al. (2016), JAMA Psychiatry
The Woman Behind the Numbers
Deborah Hasin earned her PhD in epidemiology from Columbia in 1986 and never left. Over four decades, she has published more than 550 papers on substance use epidemiology. She directed the NIDA-funded Substance Abuse Epidemiology Training Program at Columbia — training the next generation of researchers who would study the very trends her data revealed.
Her authority on this topic is not just academic. Hasin was the text editor for the Substance Use Disorder sections of DSM-5-TR — meaning she literally wrote the diagnostic criteria that define cannabis use disorder. She served as President of the College on Problems of Drug Dependence (CPDD) in 2023-2024, the oldest and largest organization dedicated to substance use research.
When Hasin publishes cannabis epidemiology data, she is not an outside observer. She is the person who defined the measurements, validated the instruments, and trained the researchers. This gives her work unusual authority — and unusual responsibility to get it right.
What This Means for You
The practical implication of Hasin's data is straightforward: if you use cannabis, there is a meaningful probability — roughly 1 in 3 among current users — that your use pattern meets the clinical criteria for a use disorder. This does not mean you are "addicted" in the way most people imagine addiction. DSM-IV cannabis use disorder includes a spectrum from mild (2-3 criteria) to severe (6+ criteria). Many people with diagnosable CUD would not recognize themselves as having a problem.
But the criteria exist for a reason. They capture patterns — using more than intended, failed attempts to cut back, continued use despite problems, tolerance, withdrawal — that predict real-world impairment. If you're unsure where you fall, our CUD self-assessment guide walks through the DSM-5 criteria, and our guide to whether weed is addictive provides the full scientific context.
Does this study prove cannabis is addictive?
It proves that a substantial minority of cannabis users — roughly 30% of past-year users in 2012-2013 — meet the diagnostic criteria for a use disorder. Whether you call that "addictive" depends on your definition. Cannabis produces genuine dependence (tolerance and withdrawal) and genuine compulsive use patterns. It does not produce the acute physical dependence of opioids or alcohol, and its withdrawal syndrome is milder. But 30% is not a trivial number — it's higher than alcohol's per-user dependence rate.
Why did the per-user rate go down while total cases went up?
Because the new users who entered the cannabis-using population between 2001 and 2013 were, on average, more casual. As cannabis became more socially acceptable and accessible, more people tried it or used it occasionally — people less likely to develop problematic patterns. This diluted the per-user rate. But the sheer volume of new users more than offset the lower rate, producing a net increase in total disorder cases.
Is 30% the same as a 30% chance I'll become addicted?
No. The 30.6% figure is a cross-sectional prevalence — the fraction of current users who met criteria at the time of the survey. Your individual risk depends on factors including frequency of use, age of onset, genetics, co-occurring mental health conditions, and product potency. Daily users have a much higher risk (exceeding 50% in some studies) while occasional users have a much lower risk.
How does cannabis compare to other substances?
In the original NCS data (Anthony 1994), the lifetime risk of dependence among people who ever tried each substance was: tobacco 32%, heroin 23%, cocaine 17%, alcohol 15%, cannabis 9%. The Hasin figure of ~30% among current users is not directly comparable because it uses a different denominator. Cannabis is less acutely addictive than nicotine or opioids, but more likely to produce a use disorder than most people assume.
Prevalence of Marijuana Use Disorders in the United States Between 2001-2002 and 2012-2013
Hasin DS, Saha TD, Kerridge BT, Goldstein RB, Chou SP, Zhang H, Jung J, Pickering RP, Ruan WJ, Smith SM, Huang B, Grant BF (2015) · JAMA Psychiatry
Related Research
Key studies in this area
Comparative Epidemiology of Dependence on Tobacco, Alcohol, Controlled Substances, and Inhalants
Anthony et al. (1994)
Cannabis Users Are Less Obese
Le Strat & Le Foll (2011)
The Dunedin IQ Study
Meier et al. (2012)
CBD Reduces Heroin Craving
Hurd et al. (2019)
What the researchers found
Past-year cannabis use rose from 4.1% of adults in 2001-2002 to 9.5% in 2012-2013. Past-year cannabis use disorder also increased, from 1.5% to 2.9% of adults. Among people who used cannabis in the past year, the proportion meeting DSM-IV criteria for abuse or dependence declined from 35.6% to 30.6%. In plain terms, more adults used cannabis and more adults had a disorder overall. But an average individual user in 2012-2013 was less likely to meet disorder criteria than an average user in 2001-2002. Increases in both use and disorder were detected across most demographic groups.
Why it matters
As laws and social acceptance shifted in the 2000s and early 2010s, this study provided a clear national snapshot of two linked trends: many more adults reporting cannabis use, and a larger number of adults meeting criteria for a use disorder. It also showed that the likelihood of disorder among users went down, pointing to population expansion as the main driver of the higher disorder count.
The numbers in context
- Past-year use: 4.1% in 2001-2002 vs 9.5% in 2012-2013. Roughly 1 in 25 adults then, about 1 in 10 later (P<.05).
- Past-year cannabis use disorder (DSM-IV): 1.5% vs 2.9%. About 1 in 67 adults then, about 1 in 34 later (P<.05).
- Disorder among past-year users: 35.6% vs 30.6%. About 1 in 3 users met criteria in 2012-2013.
- Samples: 43,093 adults in 2001-2002 and 36,309 in 2012-2013, both nationally representative.
How the study worked
This study analyzed two nationally representative, face-to-face household surveys of U.S. adults: NESARC in 2001-2002 (N=43,093) and NESARC-III in 2012-2013 (N=36,309). Past-year cannabis use and DSM-IV cannabis use disorder (abuse or dependence) were assessed in both waves to enable comparison. Estimates were weighted to represent the U.S. adult population. The design is repeated cross-sectional, so it tracks population changes over time but does not follow the same individuals.
Who was studied
N=79,402 US adults from two nationally representative surveys conducted in 2001-2002 and 2012-2013
What this study cannot tell us
Repeated cross-sections track populations, not individuals, so they cannot identify causes. Both surveys relied on self-report in face-to-face interviews, which can be affected by willingness to disclose that may have changed with shifting attitudes and laws. Although DSM-IV criteria were applied in both waves to aid comparability, differences in survey context and participation could influence estimates. Findings apply to adults only, not adolescents. The study period ends in 2013, before the more recent expansion of legal markets and product types.
How to read the evidence
Rated moderate: very large, nationally representative samples with consistent DSM-IV criteria across waves. However, estimates rely on self-report in repeated cross-sections, and changes in disclosure or survey context may influence trends.
When this study was published
Published in 2015 using data through 2013. The legal landscape, product potency, and DSM-5 diagnostic criteria have changed since, so contemporary prevalence could differ.
The bigger picture
The pattern fits a familiar public health dynamic. When a behavior becomes more common, the absolute number of people with problems related to that behavior often rises, even if the average risk for each participant declines. This study is about prevalence, not cause. It does not show why use increased or why users in 2012-2013 were somewhat less likely to meet disorder criteria. Changes in legal status, perceived risk, product availability, and willingness to report use could all be part of the story, but this design cannot separate those factors.
Replication
Not stated in abstract.
Funding
Not reported in abstract.
Conflicts of interest
Not reported in abstract.
Questions still open
- Did these trends continue after 2013 as more states changed cannabis laws and products evolved?
- How do estimates look under DSM-5 criteria for cannabis use disorder, which differ from DSM-IV?
- Which subgroups saw the largest absolute rises in disorder counts, and how do access, potency, or marketing track with those changes?
- How much of the increase reflects greater willingness to report use versus true changes in behavior?
Common questions
Did the risk of disorder for an individual user go up?
What does 'cannabis use disorder' mean here?
Were increases confined to certain groups?
Read the original research
Prevalence of marijuana use disorders in the United States between 2001-2002 and 2012-2013
JAMA Psychiatry, 72(12), 1235-1242
JAMA Psychiatry is a highly regarded, peer-reviewed medical journal focusing on psychiatric research.
Citation
Hasin, Deborah S.; Saha, Tulshi D.; Kerridge, Bradley T.; Goldstein, Rishe B.; Chou, S. Patricia; Zhang, Haitao; et al.. (2015). Prevalence of marijuana use disorders in the United States between 2001-2002 and 2012-2013. JAMA Psychiatry, 72(12), 1235-1242. https://doi.org/10.1001/jamapsychiatry.2015.1858
Explore the wider topic
- Physical vs Psychological Cannabis Dependence: What Science Actually Shows
- The Gap Between Cannabis Perception and Science: Why What We Believe Outpaces What We Know
- Cannabis Use Disorder: Am I Addicted? Self-Assessment Guide
- Cross-Addiction After Quitting Weed: When One Habit Replaces Another
- Is Weed Addictive? What the Research Actually Shows
- Is Cannabis Addictive? What the DSM-5 and Brain Science Say
- Quitting Weed and Alcohol Together: What to Know About Dual Cessation
- Is Rehab Necessary for Weed? An Honest Assessment
- Signs You May Have Cannabis Use Disorder: An Honest Self-Assessment
- Weed Vape Pen Addiction: Why Vaping Makes It Harder to Quit