Research has not shown that marijuana causes people to use harder drugs: most people who use cannabis never move on to other illegal drugs, and much of the link between cannabis and later drug use is explained by shared risk factors such as early mental health problems, delinquency and easy access to drugs. Cannabis use is still statistically associated with later drug problems, especially when it starts in the early teens and is heavy, and some animal studies suggest a biological effect. The honest verdict is “associated, not proven causal.”
The gateway theory is one of the oldest arguments in drug policy. It holds that using a “soft” drug like marijuana opens the door to “hard” drugs like cocaine and heroin. This page walks through what the main studies actually found, as of October 2026, including the 2006 University of Pittsburgh study that our original article covered, and explains why scientists still argue about it. For background on cannabis and the brain, see our cannabis research hub.
Evidence at a Glance
| Claim | Evidence level | Source |
|---|---|---|
| Cannabis use is linked to developing dependence on other substances (alcohol, tobacco, illicit drugs) | Moderate (statistical association) | National Academies 2017 |
| Cannabis use changes rates and patterns of other drug use | Limited | National Academies 2017 |
| Cannabis use leads to starting tobacco | Limited | National Academies 2017 |
| Most people who use cannabis go on to other illegal drugs | Not supported; NIDA says most do not | NIDA |
| The usual “alcohol and tobacco, then cannabis, then other drugs” order reflects common causes, at least in part | Moderate (large cross-national surveys and prospective studies) | Degenhardt 2010, Tarter 2006 |
| Adolescent THC exposure increases later heroin intake | Preclinical only (rats) | Ellgren 2007 |
What Is the Gateway Theory?
The gateway theory describes a common order in which people start using drugs: alcohol and tobacco first, then cannabis, then other illegal drugs such as cocaine, stimulants or opioids. Researchers call this the normative sequence. Nobody disputes that the sequence is common. The argument is about why it happens.
There are three main explanations, and they are not mutually exclusive:
- Causal (biological) gateway: THC changes the developing brain in a way that makes other drugs more rewarding or more addictive.
- Social gateway: buying and using cannabis puts people in contact with sellers and peers who offer other drugs, and lowers the perceived barrier to trying them.
- Common liability: the same underlying traits and circumstances, such as genetics, mental health problems, impulsivity, family stress and neighborhood drug availability, drive both cannabis use and later drug use. Cannabis comes earlier simply because it is easier to get.
The National Institute on Drug Abuse (NIDA) reflects this split. It says most people who use cannabis do not go on to use other substances, that risk factors for cannabis use are similar to those for other addictive drugs, and that cannabis use may cause brain changes that could make addiction to other drugs more likely.
The Pittsburgh Study: Gateway Order Did Not Predict Problems
Our original article covered a University of Pittsburgh study led by Ralph Tarter, published in the American Journal of Psychiatry in 2006. The team followed 224 boys from ages 10 to 12 until age 22 and compared three groups: boys who used only alcohol and tobacco (99), boys who followed the classic gateway order into marijuana (97), and boys who used marijuana before alcohol or tobacco (28).
Key findings:
- About 22% of the boys who used marijuana did not follow the gateway order at all, so the sequence is not fixed.
- Across 35 psychological, family, peer, school and neighborhood measures, no specific risk factor explained the move from legal drugs to marijuana, apart from delinquency.
- The reverse order (marijuana first) predicted a later substance use disorder just as well as the gateway order.
The authors concluded that a tendency toward deviant behavior and drug availability in the neighborhood drive marijuana use, which supports the common liability model rather than a drug-specific gateway. The study was small and limited to boys in one region, so it is one piece of evidence, not the final word.
Studies That Support a Gateway Effect
Several well-designed studies found that the link survives attempts to explain it away.
- Australian twin study (2003). Researchers studied 311 pairs of twins in which one twin used cannabis before age 17 and the other did not. The early user had 2.1 to 5.2 times higher odds of later drug use, alcohol dependence and drug abuse or dependence than the co-twin. Because twins share genes and family environment, shared background alone could not explain the gap. The authors pointed to the social context in which cannabis is bought and used as a likely reason.
- New Zealand birth cohort (2006). A 25-year study of 1,265 children found that more frequent cannabis use was tied to use of, and dependence on, other illicit drugs, even after statistical control for many confounders. The association was strongest in adolescence and faded quickly with age. The authors said the mechanism remained unclear.
- US national survey (2015). Using data from more than 6,600 Americans who used cannabis before any other drug, researchers estimated that 44.7% went on to use another illicit drug at some point in life. Mental health disorders and more severe substance use predicted who progressed.
- Rat studies. Adolescent rats given THC self-administered more heroin as adults and showed changes in the brain’s opioid system. This supports a possible biological pathway, but rat findings do not prove the same effect in people.
Studies That Point to Common Causes Instead
- RAND modeling study (2002). Researchers built a model of US drug use in which marijuana had no causal effect on hard drug use, only a shared “propensity” to use drugs. The model reproduced every pattern usually cited as gateway evidence. The authors said a gateway effect may exist, but the data do not require one.
- WHO World Mental Health Surveys (2010). Comparing 17 countries with very different drug use levels, researchers found the gateway order varied with how common each drug was locally. They concluded the pattern at least partly reflects common causes, and that preventing use of “gateway” drugs might not by itself reduce use of later drugs.
- US National Comorbidity Survey Replication (2009). In more than 9,000 adults, people who broke the usual order were mostly no more likely to become dependent. Mental health problems that began early in life predicted both unusual drug sequences and dependence.
What the National Academies Concluded
The 2017 National Academies of Sciences, Engineering, and Medicine report remains the most thorough official review. On other substances it found:
- Moderate evidence of a statistical association between cannabis use and developing dependence on, or a use disorder with, other substances, including alcohol, tobacco and other illicit drugs.
- Limited evidence that cannabis use is associated with starting tobacco use.
- Limited evidence that cannabis use is associated with changes in rates and patterns of other drug use.
“Statistical association” is the key phrase. The committee did not conclude that cannabis causes later drug use. It also found substantial evidence that starting cannabis at a younger age and using it more often are risk factors for problem cannabis use, which matters more for most people than the gateway question.
Who Is Most at Risk
Across the studies above, the same warning signs keep appearing:
- Starting young. The association with other drug use is strongest in the early and mid teens. NIDA notes that using cannabis at a younger age raises the chance of developing a cannabis use disorder.
- Frequent or heavy use. The New Zealand study found that frequency, not just any use, tracked with other drug use.
- Mental health problems. Early anxiety, depression and behavior problems predicted both progression and dependence in the US surveys.
- Delinquency and drug availability. The Pittsburgh study tied marijuana use to deviant behavior and neighborhood access.
If cannabis use is becoming hard to control, our guides to cannabis use disorder and treatment options cover the signs and what helps.
Risks and Side Effects
Setting the gateway debate aside, cannabis has documented risks of its own. NIDA describes cannabis use disorder as a real and common outcome of regular use, and the National Academies found substantial evidence that earlier and more frequent use raises that risk. For teenagers, whose brains are still developing, the case for waiting is strongest. See our pages on cannabis and the developing brain and cannabis and IQ for that research.
Talk to Your Doctor
If you are worried about your own use, or a teenager’s, a primary care doctor or addiction specialist can screen for cannabis use disorder and for the mental health conditions that often travel with it. Treating depression, anxiety or ADHD early can lower the overall risk of problem drug use. Be honest about everything you use, including alcohol, nicotine and prescription drugs, so you get advice that fits.
Gateway Theory FAQ
Is marijuana a gateway drug?
Not in the simple sense of causing people to use harder drugs. Cannabis use is statistically linked to later use of other drugs, but most users never move on, and shared risk factors explain much of the link. The National Academies rated the association moderate and did not call it causal.
What percentage of marijuana users go on to harder drugs?
Estimates vary by study and country. One US national analysis projected that 44.7% of people who started with cannabis would use another illicit drug at some point in life, which also means most did not. NIDA states that most people who use cannabis do not go on to use other substances.
Who came up with the gateway theory?
The idea grew out of drug-use research showing that people tend to start with alcohol and tobacco, then cannabis, then other drugs. The studies on this page test whether that order reflects cause or common risk factors.
Are alcohol and tobacco gateway drugs too?
By the same logic, yes. In most people’s histories, alcohol and tobacco come before cannabis, and the National Academies found limited evidence that cannabis is linked to starting tobacco. The WHO surveys found the order depends on which drugs are most available locally.
Does legal cannabis increase other drug use?
The evidence is not settled. The National Academies found only limited evidence that cannabis use changes rates of other drug use, and we have not found a definitive review that answers the question for legal markets.
Does starting marijuana young matter?
Yes. Early use is the most consistent warning sign. The National Academies found substantial evidence that starting earlier raises the risk of problem cannabis use, and the gateway link itself is strongest in adolescence.
Adult-use cannabis is legal only for people 21 and older in states such as Washington and Oregon. For more plain-English explainers, see Cannabis 101.
For adults 21+ in legal states. General information, not medical advice. Talk to a doctor before using cannabis for a health condition.
