- A major US survey finds teens who smoke cigarettes are more likely to vape than vice versa, directly challenging the “gateway theory.”
- The widely-cited idea that vaping causes smoking may be a misinterpretation of limited data, researchers say.
- Vaping and smoking are both more closely linked to other risky behaviours than to each other.
- Study suggests prevention strategies for teens should focus on wider behavioural patterns.
A large national study of American teens has found smoking is more likely to lead to vaping than the other way around – challenging the theory that vaping acts as a “gateway” to smoking.
The research, published in the journal Preventive Medicine Reports, used data from around 7,700 US high school students and compared vape and cigarette use with a broad range of other behavioural and mental health risks.
While vaping and smoking were correlated, researchers found that each was more strongly associated with different behaviours – vaping with alcohol and cannabis use, and smoking with other tobacco products and illicit drugs.
This suggests that both behaviours may stem from shared risk factors, not from one directly causing the other.
“Our inferential results show that the narrow unidirectional relationship of interest… is not unique,” the authors wrote, “in that the reverse-direction models produce very similar effect sizes.” In fact, they found that cigarette smoking was a stronger predictor of vape use than the reverse.
Challenging the “gateway” narrative
The new findings directly counter the so-called gateway theory. This is the idea that vaping inevitably leads to smoking, especially in young people. This theory has underpinned much of the public discourse and many policy decisions around youth vaping. But the authors say it may rely too heavily on narrow, two-variable models that don’t reflect the complex web of teenage behaviour.
“Much research on youth risk behaviors has a narrow focus on two to three variables and a causal-inference framework,” the study notes. “This can result in biased interpretations when there are many correlated risk behaviours.”
By using a broader ‘predictive-inference’ framework – designed to assess how well one behaviour predicts another, regardless of assumed causality – the study shows that vape use is not the reliable smoking predictor it is often made out to be.
The researchers found that ‘reverse-direction models’, which looked at how smoking predicted vaping, actually performed better at classifying youth behaviour than models assuming vaping caused smoking.
This kind of statistical modelling is rare in public health research but allows a more realistic view of adolescent behaviour, according to the authors.
Looking at the wider picture
The study also employed three exploratory data methods – Spearman correlation, multidimensional scaling, and hierarchical clustering – to examine how vape and cigarette use grouped with other behaviours. It found that vaping clustered more closely with alcohol and cannabis use, while smoking was more associated with other tobacco products and illicit drug use.
This supports a “common liability” theory of substance use, in which young people who are more inclined toward risky behaviours may try both smoking and vaping, without one necessarily causing the other.
The authors wrote: “Exploratory data analyses demonstrated that a narrow focus on two variables at a time… overlooks other, more strongly correlated risk behaviours.”
In short, youth who vape are not necessarily future smokers. They are more likely to be youth who are already experimenting with substances more broadly.
Implications for prevention
The authors argue that youth health strategies should avoid singling out vaping as a leading risk factor and instead focus on wider behavioural patterns. For example, preventing youth smoking might be more effective by identifying students who also use cigars or illicit drugs rather than those who vape.
Similarly, they warn against oversimplified interpretations of vaping’s link to mental health. While other studies have suggested that vape use may lead to depression or suicide risk, the evidence shows that mental health symptoms typically precede substance use, not follow it.
“If reducing suicide is the goal, it would be more effective to screen adolescents for self-reported sadness or hopelessness,” the study notes.
A call for smarter science and policy
The research ultimately calls for a more sophisticated approach to understanding youth behaviour – one that uses comprehensive data and predictive models rather than narrow assumptions of cause and effect.
“Cross-validation testing to examine generalised classification performance extends the standard causal-inference practice… to a complementary predictive-inference framework which is uncommon in this field,” the authors wrote.
While vaping among teens remains a concern, the study offers reassurance that fears of it causing a smoking epidemic may be misplaced. And for adult smokers considering switching to vapes, it reinforces the case that vaping is a safer alternative, not a slippery slope.
