Science Lesson: Does Vaping Lead To Smoking?
The “gateway effect” refers to a claim that one disparaged behavior – usually a relatively harmless one – causes people to later take up a second, more harmful disparaged behavior. (“Causes” means that they would not have taken up the second behavior absent the first behavior, but would do so if exposed to the first behavior). The concept was concocted during the early War on Drugs to provide an excuse for punishing people for using cannabis based on the claim that it would cause them to turn to “hard drugs.” More recently the claim has made, without evidence, about smokeless tobacco use causing smoking. And now it is used to suggest vaping leads to smoking.
Almost every claim you ever read about vaping being a gateway to smoking – usually focusing on teenagers – is nonsense. This is common knowledge among vapers. However, you might not fully realize why.
The most important point to understand is that quantities matter. Any claim about a gateway effect that does not include numbers is meaningless. Claiming “there is a gateway effect” might mean “30 percent of would-be nonsmokers who try vaping take up smoking within one year.” Or it might mean “5 percent of would-be nonsmokers who are regular vapers become smokers within five years” or even “a single person became a smoker because of e-cigarettes.” These are obviously very different claims, and so the generic phrasing is propaganda, not a meaningful statement.
Evidence shows the first of those possible interpretations is false. No study has ever come near to suggesting that 30 percent of new vapers take up smoking within a year. The second is scientifically plausible and research could be designed to assess it. That said, legitimate research has never been done, so there is no basis for claiming it. The third is obviously true, but statistically trivial. There have been many chances for a gateway effect to happen to one person and are many paths via which it could happen, and so it is implausible that it has never happened. Each of these very different claims describes a gateway effect. Unless anyone claiming there is a “gateway effect” clarifies what they mean, it is disingenuous and factually incorrect to say there is any such thing as “the” gateway effect.
Since unquantified claims about a gateway effect are meaningless – that they could refer to something that is obviously false, something that is obviously true, or something uncertain in between – gateway accusations coming from vaping opponents are meaningless. Not a single such claim (to my knowledge) has included proper quantification. Vaping opponents never say “this study suggests that 6 percent of never-smokers who become regular vapers start smoking within two years.”
The second problem is that none of the existing studies could possibly detect a gateway effect. Several are so bad that they simply observe whether people have both vaped and smoked. They cannot distinguish between gateway cases, smokers who took up vaping to quit smoking, and people who tried both but neither caused the other. It is impossible to know whether there are any gateway cases among the study subjects because there will certainly be some in each of the other two categories.
Slightly better studies identify the order of the behaviors, focusing on subjects who vaped before they ever smoked. But gateway does not refer to mere ordering, but causation. It is inevitable that more vapers will take up smoking than non-vapers because they have already demonstrated they like consuming nicotine (about half the population does not), that they are inclined to engage in a socially disparaged behavior, and – in the case of the minors who are the focus of the claims and studies – that they are willing to commit a status offense.
Tobacco controllers and other drug warriors call this the “common liability model” because they cannot pass up a chance to create derogatory vocabulary. Scientists call it confounding, specifically confounding due to a common cause. Confounding means there are differences in the outcome (smoking) between the exposed population (vapers) and the unexposed population that are not caused by the exposure. In this case, those underlying individual characteristics cause someone to be inclined to vape and also cause her to be more inclined than average to smoke. Other sources of confounding further complicate the analysis.
Not one study of vaping has made any serious attempt to deal with confounding. There was one study of smokeless tobacco that did. It found that any apparent gateway effect, supposedly observed in other studies, disappeared. You can read more about this and how a legitimate attempt to detect a vaping gateway effect could be done in my paper on that topic.
In short, there is literally no study evidence that shows a vaping gateway effect. This is even apart from the fact that the claims are unquantified.
However, most of the claims by vaping proponents that there is no gateway effect are also nonsense. These claims are often based on observing that there are few never-smokers who start vaping. While this is true, it merely means that the population at risk of a gateway effect (never-smoking vapers) is small. This means that even if there is a modest gateway effect, the total number of gateway cases will be small. But it does not address whether there is a gateway effect.
Similarly, some claims are based on the observation that smoking rates among particular populations (e.g., American teenagers) have decreased as vaping has become popular. But this merely means that any gateway effect is small enough that it is dwarfed by the (unknown) effect of vaping causing a reduction in smoking plus the (unknown) effect of other factors that are lowering the smoking rate. A gateway effect could still be hidden there.
Indeed, any claim that there have never been any gateway cases is intuitively false. Of course there are would-be nonsmokers who took up smoking after discovering vapes. And while it is possible to design good, quantitative studies to measure any gateway effect (even if those have yet to be conducted), any claim that a particular study shows there is no gateway effect is incorrect. All we can conclude from a study (no matter how good) is that any effect was too small to detect using those methods in that population. It might have been detected if different methods were used and might exist in another population. This is a universal challenge in science: If you fail to observe something it might be because it does not exist, but it might be because you did not look for it correctly.
Finally, it is useful examine the unstated nonsensical assumption that underlies gateway claims: “of course, it is only natural that people will ‘graduate’ to harder drugs.” In reality, why would someone who is happily using nicotine in a low-risk form decide to switch to a high-risk product? It is possible, of course, but a sensible assessment suggests it is not natural, but actually quite odd.
Indeed, the evidence we have shows that it is sufficiently odd that we have not been able to see it happening. A large effect would be detectable, even with relatively poor scientific methods, but no such effect has been detected. Vaping opponents have not produced even a single testimonial from someone who claims that she never would have considered starting smoking, but vaping caused her to do so.
You can find longer versions of my analysis of the vaping gateway effect in this blog post and still more here.