“E-cigarettes are a gateway” is a genuine scientific claim

by Carl V Phillips

The latest entrant into the Dunning-Kruger gateway follies is ASH Wales, with this report that is headlined, “New research shows e-cigarettes are not a gateway for young people to take up smoking”. What evidence do they present that supports this remarkable claim that a universal negative is true? None.

The gateway claim, that using e-cigarettes causes some would-be never-smokers to smoke, is a legitimate scientific hypothesis. As such, it should not be asserted to be true (to a nontrivial extent) without useful evidence, especially since it is such an unlikely causal sequence, as I have explained elsewhere. The assertions that gateway effects are occurring have been based on evidence that does not actually show that. This is certainly the major problem in this area. But similarly, the claim should not be dismissed with word games or junk science. In this particular case, ASH Wales — like many others before them — seem to not understand a 101-level point from epidemiology, the difference between “not many people are at risk” and “it never happens among those who are at risk.” They claim that because a large majority of e-cigarette users among teens have already smoked, there is therefore no gateway effect. Um, yeah. 

Even setting aside the technical point that someone who has already smoked can become a gateway case, this is obvious nonsense. The gateway claim is never that most people exposed to X are caused to do Y as a result, but that some are. The fact that gateway proponents never attempt to quantify “some” — to estimate what portion of those who are at risk do become gateway cases — is a good reason to doubt they are really attempting to do science. But it also seems to confuse their opponents.

ASH’s logic is akin to saying, “most people in prison are already inclined to be criminals, and therefore the experience of imprisonment does not increase the probability someone will go on to engage in a life of crime.” Or to make it more obvious, though a bit less analogous, “over 90% of the subjects in our study were passing high school, and therefore special tutoring programs cannot help students pass high school.” That observation could be used as argument to not bother funding those programs, of course, but is obviously silent on the question of whether they work. Similarly the supposed evidence used by ASH (and many others who object to the gateway claim) is silent on their claim.

The question is not whether many people are immune to the effect (“not at risk” in the language of epidemiology) because they already have the outcome. It is whether those who are at risk are ever caused to smoke. More precisely, it is how many of those who are at risk of being gateway cases become gateway cases. As I noted in the previous post, it is basically certain that there will be someone fitting that description. Even if it so happens this has never been true, due to some wildly improbable stroke of luck, it obviously could become true tomorrow (thus the humorous  innumeracy of ASH’s claim of the universal negative).

Another line of nonsense (nonscience?) deployed by gateway claim opponents is that because the net effect of e-cigarettes is less smoking in this population, there is no gateway. Wrong again (and not because of the empirical tenuousness of that claim). Whatever the effect on smoking cessation or prevention, it does not mean there is not also a gateway effect; indeed, it might be entirely independent of it. The word games of saying “it is a gateway away from smoking, ha ha nyah nyah” are particularly inappropriate. The jargon has a particular meaning, and that is not it. Trying to twist a word, as if that constituted an argument against the claim it was used in, is marketing rhetoric not scientific analysis.

Why does this nonscience happen? I am increasingly convinced it is caused by the notion on the part of opponents of the gateway claim (to say nothing of its champions) that if there is a gateway effect — any gateway effect — then some draconian intervention is warranted. So they feel a need to make the claim that there is absolutely no gateway effect, something that is both absurd to claim (being a universal negative) and undoubtedly false. In the previous post I pointed out that this is bad tactics on the part of e-cigarette proponents, conceding every step in the missing middle of the series {empirical result, …, policy recommendation}. But for many of them it is not just bad tactics, it is fundamentally bad reasoning. They come from the public health mindset that treats an empirical result as being tantamount to a recommendation of a particular policy.

Someone could argue that so long as the gateway effect is small enough (e.g., that it is so rare that there is no empirical evidence it exists), it does not justify policies that interfere with adults’ choices. Or they could go further argue that teenagers, like adults, should also be free to choose. Or someone could argue, if they believed in the public health pseudo-ethic, that the net effect of e-cigarettes on teenagers is reduced smoking, and therefore nothing should be done that interferes with this, such as banning sales to minors. (An implicit corollary to this is that if it stops being true, and the gateway effect starts to exceed the smoking prevention effect, then teenagers should be stopped from using e-cigarettes to the extent possible.) On the other hand, someone could instead argue that whatever the net effect is, those individuals who would be gateway cases should be protected from the bad outcome caused by e-cigarettes (for those keeping track at the level of ethical rules: this would be based on a combination of a particular view of “trolley problem” ethics, the assumption that being a gateway case is bad, and the notion that the availability of e-cigarettes should be treated as the act of commission and the act of stopping it should be treated as the omission or default state — it’s a complicated and somewhat tortured but not utterly unreasonable position). More simply, someone could argue that a particular proposed policy would not actually reduce whatever gateway effect exists.

The general point is that “what should we do?” is a very different question from “what is the best scientific assessment of what true?” But if someone does not understand that rather obvious point, and acts as if evidence of a particular worldly outcome is tantamount to a policy recommendation, then they are forced to try to play games with the evidence in order to defend the particular policy they prefer. Notice that this means that someone prefers a particular policy regardless of what the evidence really shows. It truly baffles me that this can be the view of so many in public health, and yet they still do not understand that there must be some steps in between the evidence and the policy preference. Even more baffling, they still fail to understand this when confronted with the observation that others prefer a different policy. Just where do they think political opinions come from?

It is apparent that the utter blindness about ethics and political disagreements plays no small role in contributing to the denigration of science in public health.

This was exemplified in a Twitter exchange I had with ASH about their report (in which, incidentally, they did not even attempt to defend their reasoning). For those who do not know, ASH-Wales is generally pro-ecig in their political statements, though not so much in their actual actions (this disconnect was referenced by others in the Twitter exchange). So having taken a political position, they are then motivated to produce the nonscience to support it. Why? Because they do not even understand they are engaged in political advocacy. When I noted, to a third party in that exchange, that the ASH report was basically uninformative but the headline was informative about ASH’s political goals, ASH’s tweeter expressed what I believe was genuine astonishment at ASH being referred to (obviously accurately) as a political organization. That would be the problem with most political activists in public health, including those on both sides of the e-cigarette debate. How can someone assess the ethical basis of their political position when they do not even understand they are asserting a political position?

(Is this like Molière’s “speaking in prose all my life”? Or is it more like a cartoon version of a religious true-believer who cannot even understand he is believing something rather than merely doing the only possibly right thing? Perhaps it is like a critter from Aesop who simply has no notion of how to respond to the question of “why do you do that?” because that is simply what that animal does. I am still working on this.)

Finally, to circle back and sum up, all of the following are true (and thus obviously do not contradict one another as is sometimes implied):

  • There is no empirical evidence that strongly suggests there has been so much as a single gateway case.
  • It is impossible to show there have been no gateway cases, let alone that there never will be.
  • The gateway causal pathway seems extremely unlikely given the particulars of the situation.
  • Sheer number mean there almost certainly has been at least one gateway case.
  • The fact that only a small portion of e-cigarette users are at risk of being gateway cases (for the moment) does not mean there is no gateway effect, and does not even affect the probability of it for a given at-risk individual in any obvious way.
  • Even if there is a gateway effect as large as is conceivable, this is not sufficient to justify any policy.
  • Even if there is only a trivial gateway effect, this is not sufficient to justify not pursuing policies to reduce the gateway effect (setting aside that there has been no policy proposed that seems like it would further that aim).
  • Even if there are clearly more people who are prevented from smoking than caused to smoke via a gateway effect, this is not sufficient to justify precluding a policy that reduces both of those effects.

Anyone whose scientific or ethical claims are based on denying one of these points is probably trafficking in nonsense.

18 responses to ““E-cigarettes are a gateway” is a genuine scientific claim

  1. You have basically said what everybody means, others are simply not as detailed and literal. The “Public Health” control freaks are not running around saying “there is a teensy-weensy gateway effect”, when they say a gateway effect they are implying, by intonation and phrasing, that it is a large effect and essentially a NET effect of causing more kids to smoke than those it prevents. When we say “there is no gateway effect” we are saying “there is no effect such as you are implying.” Your points at the end say the same thing to anybody who isn’t a nit-picker. (BTW, I am a nitpicker at work, just not after I get home. And my work is with computers, and you would not enjoy the caveats I make when on my own professional topic.)

    • Carl V Phillips

      Sorry, but no. Ecig cheerleaders are most definitely saying there is *no* gateway effect. It is just as easy to say that the gateway effect is smaller than the smoking prevention effect, but they are not saying that. Perhaps they do not understand what they are saying (the Dunning-Kruger point), but there is no grounds for claiming that what they are very clearly stating clearly means something different. There are cases like that in the world, of course, where the meaning of a sloppy shorthand is fairly clear. This is not one of them.

      Exactly what gateway theory proponents are claiming is a more substantive question. They are indeed suggesting that the effect is big enough to warrant some policy. I do not read what they are claiming as there being a net increase in smoking right now. However, the utter failure to try to quantify, even when they report empirical results (except to the extent that they disingenuously imply the entire association they observe is a gateway effect) is, as I said, a clear indication that they are not even trying to do science.

      • I’ve sat in more than a dozen of those hearings and listened to more online. I know what it sounds like, and how it is heard by the legislators listening to them.

  2. I don’t get your very last point. A policy that prevents both stopping millions of people from quitting and hundreds from starting smoking should not be precluded?

    • Carl V Phillips

      There are two easy answers to this that come to mind, and I suspect many more could be added with a bit of thought:
      1. Someone posits that this is a temporary state of affairs and that the balance will change over time, so intervention should take place sooner rather than waiting until after that.
      2. What I mentioned in the post, about the trolley problem. There is an ethical case to be made that affirmative action that hurts one person who would have otherwise been unharmed cannot be justified in order to save five others (or whatever ratio you want to use) from an act of fate. (Consider: would you kill one healthy person and harvest his organs to save five others who needed them?) If one chooses to see the supplying of ecig as the affirmative act (as opposed to seeing the removal of them as the affirmative act), as many in public health do, then it could be argued that causing harm to the poor gateway case individuals is unethical, even if it helps even more of their peers.

      Being able to identify these arguments is not the same as saying they are right. But just ignoring them, as if they can be dismissed without even addressing them (as is typically done by those who do not find them compelling), is not the same as arguing that they are wrong.

      • OK, basic parsing problem. The “both” confused me, it looks like it means preventing saving anybody regardless of whether or not it would cost other lives, i.e. as in there is no reason to try to stop the the trolley crash by a means that WON’T kill someone else.

  3. The main point thing is that they’re not a gateway to smoking in an actual sense, and I think that is what’s meant. It’s unlikely that for those who already smoked ecigs would do what cigs have not.

    • Carl V Phillips

      Um, I am really not sure what you are saying. If you are saying that it is simply impossible that a would-be never-smoker could ever become a smoker as a result of e-cigarettes, this is obviously wrong. It seems unlikely in any given case, for reasons I have spelled out, and thus we would want to see some very convincing evidence to believe it happened much, but that is quite different from impossible.

      • 1. Sorry, I meant to write they’re not an ACTIVE gateway to smoking.
        I mean, if someone went around injecting dilute e-liquid into kids, that may be a gateway causing future smoking, but that’s not how e-liquid is being used.

        2. Question: If using e-cigs lower nonsmokers’ chance of future smoking, but also cause some nonsmokers to take up smoking, is that considered a gateway effect?

        • Carl V Phillips

          It does not matter what the intermediate steps are. If using/trying e-cigarettes causes someone to smoke then it is a gateway. So, yes, even if it prevents more smoking uptake than it causes, there is still a gateway effect. Compare: Having gun for self-protection does sometimes result in successful self-protection, despite the fact that it is about 20 times more likely to kill the owner or a member of his family. Or: Vaccines save a lot of lives; they also occasionally kill the recipient.

  4. “It truly baffles me that this can be the view of so many in public health, and yet they still do not understand that there must be some steps in between the evidence and the policy preference.”
    Why does this baffle you? I think of public health as a ‘form of life’ and like all forms of life it requires agency to perpetuate itself. Acknowledging the steps in between the evidence and policy preferences would render it impotent.
    The capacity of public health to act in the world depends on its privileged status as experts. Whether this status is deserved or not is another argument. But the fact is whether it can bring about a change in the world (its preferred policy) depends enormously on its status as experts in the field of health (biological welfare). The influence it has to bring about change (agency – to act in the world) depends on this perceived expertise.
    However, the moment you recognize that human welfare cannot be reduced to the question of biological welfare (public health area of expertise) you have removed their perceived status as experts along with capacity to act in the world. This is I guess why we believe in freedom and democracy – because we recognize that however fallible the individual may be, they are nonetheless in the best position to judge questions of their own welfare (in human rather than biological terms).
    But to recognize this is to recognize that public health has no privileged position in these questions and hence their capacity to act and bring about change is stymied and curtailed. Indeed, they would become mere chroniclers of risk (biological risk) without the capacity to act and bring change in the world.
    To me the pseudo ethic at the heart of public health thinking is not so much a fault or limitation in their thinking, but rather seen from within its own form of life an essential feature. Its capacity to act depends on an active ongoing suppression. In short to quote one of my favourite philosophers:
    “Regard any morality from this point of view: it is ‘nature’ in it…which teaches the narrowing of perspective, and thus in a certain sense stupidity, as a condition of life and growth.” Nietzsche BGE 188

    • Carl V Phillips

      This could be seen as something of a case for why they might want to *act* as if there were no intermediate steps between empirical observation and policy recommendation. But most of them seem to genuinely not recognize that there are intermediate steps that they are not acknowledging. It would not necessarily be baffling for people to pretend to believe something that is glaringly obviously wrong in pursuit of their self-interest. It is baffling when it appears they are not pretending.

  5. “But most of them seem to genuinely not recognize that there are intermediate steps that they are not acknowledging.”
    I read this fact as evidence of how important suppressing this acknowledgment is to maintaining their form of life and the agency which is essential to it. I imagine this suppression is largely unconscious, an instinctive resistance to that which threatens it.
    People think of self delusion as being destructive, and whilst it obviously can be, it would not exist if it did not have survival benefits.

  6. Yes, someone somewhere, a youth who has never been the slightest bit interested in tobacco cigarettes, will experience a ‘puff’ on an ecig and obtain an ecig and puff away. As a result, the youth will become aware of tobacco cigarettes and obtain a packet, and ‘transition’ from ecigs to tobacco cigs. It will happen, and it might well happen quite frequently.
    The crux of the matter is what does ‘quite frequently’ mean in population terms?
    There is a reasonable case for a certain amount of impression to be acceptable. For example, all around the world, the RSP claim that ecigs are ‘95% less dangerous’ has been quoted again and again and again. But, that phrase has mutated into ‘at least’ 95% less dangerous.
    Will claims that ecigs are definitely not gateways to smoking also mutate a little? Will they mutate into ‘extremely unlikely to be’ gateways into smoking? For that is what the claim should have been in the first place.
    But it seems that, in order to get the publicity, in the first instance, the claim must be ‘over the top’. In that case, “Public Health” has to find an individual case where a youth transitioned from ecig to tobacco cig, or contest the claim of perfection in some other way. Even if “Public Health” does not contest the claim that ecigs are definitely not a gateway to smoking, the claim can be ‘modified’ over time as I have described.
    I have read a little of the book ‘Science on the Verge’ and intend to read the rest asap.
    I feel that there is something wrong with the definition of ‘Science’. I know that, in scientific circles, there are such things as ‘hard science’ and ‘soft science’. Erm… No. There is only one science, which is ‘hard’ science. ‘Soft’ science is not science. ‘Soft’ science needs a new name. Perhaps that name could be ‘Astrology’.

  7. Roberto Sussman

    I agree that a “gateway” effect cannot be ruled out `a priori’ and that it should be studied as rigorously as possible. Perhaps to produce theoretical models or to guide observations on this effect various external factors need to be examined and quantified, for example the relative availability/price of e-cigs and regular cigarettes. I have seen that some never smokers whom I meet have become curious enough to try vaping and (surprisingly) find it pleasant. None of them (as far as I know) have adopted the habit, but we cannot claim that no never smoker will ever take up vaping. Some might do it. Pending various factors perhaps a lot in the long run.

    Now, let’s consider the following hypothetical case scenario: a never-smoker becomes a regular vaper and truly enjoys it, but suddenly e-cigs and vaping gear become unavailable or illegal or very hard to get in the place where he/she lives. Our never-smoker-turned-vaper runs out of e-juice and vaping supplies and cannot vape, he/she misses the experience and ritual. Suddenly in a party or social gathering somebody offers him/her a cigarette (or is invited to a hooka session) and he/she starts smoking. Bingo !! Gateway !!

    In my opinion there could be reasons to infer that such scenario could be infrequent. Perhaps the transition from vaping to smoking could be much harder than from never-smoking to vaping. Vaping does not produce a lot of effects that make smoking very unattractive to non-smokers: it is in general much more gentle than smoking to their mouth, nose, throat and lungs (more so if just puffing without inhaling), the exhaled vapor rapidly disperses and leaves no noticeable trace odor, and requires no fire and produces no ash. On top of this, scenarios of this type should be influenced by relative costs (sin taxes) and hardening of regulation on vaping. I wonder if models have been devised by considering such hypothetical scenarios.

    • Carl V Phillips

      Yes. I have worked out some modeling of all that. Others claim to have, but they seems to be just producing simple deterministic equations with a fancy overlay.

      There is no doubt that a gateway can be engineered by policy, making smoking more attractive for would-be nonsmokers. This was done in Canada, driving smokeless tobacco users to smoke, as recounted on this page. My analysis about gateway effects has stuck to a relatively status quo policy situation. But serious restrictions on e-cigarettes could do the same thing. It is not perfectly clear whether you would want to define “ex-smokers who quit with a THR product but were driven back to smoking” to be gateway cases, however. Certainly they would not be if they would simple still be smokers if the other product had not existed. If they might have quit by other means and thus been unaffected by the policy change, you might call them gateway cases, though this certainly would qualify as chutzpah given that the same people who whine about gateway cases would be the ones responsible for the policy.

      Vaping may not make smoking terribly more attractive, but certainly somewhat. It is more like smoking than anything else people do, after all. The seemingly most likely candidates for gateway effects are people who would have preferred smoking to nonsmoking, but were manipulated and bludgeoned into never finding that out. This seems to be what the tobacco controllers are really concerned about — loss of their ability to control. Of course, ethical people would say that someone who prefers smoking to not smoking is made better off by the gateway effect, not worse.

      • natepickering

        Carl, I think your theorem about the “natural” rate of tobacco use (that is to say, the nominal rate of tobacco/nicotine use in a society in which 1) the most widely used products are negligibly harmful; and 2) government policy and media content not crafted by militant tobacco controllers) being around 50% is probably correct, if not even a shade on the low side.

        And I think the tobacco controllers, at least insofar as those in its leadership, believe it too. Thus, when they argue that their own efforts are the only thing preventing a massive increase in new overall tobacco/nicotine use, they’re making a perfectly accurate and fact-based assertion. The actual lie is in the (unspoken yet always strongly implied) assertion that “more tobacco/nicotine use” is automatically a disastrous public health outcome, which it quite obviously is not. If every cigarette smoker switched to smokeless/vaping tomorrow, and 75% of nonsmokers started using those products at the same time, the overall net benefit to public health would be inestimable.

        • Carl V Phillips

          Yes. In my series about why is there anti-THR, I pointed out that the anti-tobacco extremist faction in public health, whose goal is not a matter of health, but minimizing tobacco use, would consider that outcome to be terrible. And, no, they will never openly state their actions are based on their real goal/motivation because they know there is very little support for it.

Leave a Reply to natepickering Cancel reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s