by Carl V Phillips
Apologies for my ongoing fascination with Stanton Glantz. As regular readers know, I have made a minor hobby of trying to figure out what really makes him tick, and that I sometimes hesitate to call him a liar because he seems to believe some or all of the crazy things that he claims. I have called him out as a liar because at the very least he lies when he claims to understand what constitutes good and honest science. Indeed, it seems impossible that he does not know he is lying with statistics in his claims about ETS. But he might not be the calculated and knowing liar that many others are. For example, he seems to really believe his claims that smoking in movies causes most smoking initiation; and why would someone flog such an embarrassingly nutty claim and go to war against free artistic expression except out of genuine belief? In general, he may not even know enough to know that he does not understand science. But he is aggressive and somewhat influential, and that makes him fair game, so….
My latest fascination is with this little post by him, which was basically just housekeeping. The purpose of it was apparently just to explain why he deletes comments on his blog about e-cigarette success stories. I suspect he has been deluged with those since he started becoming aggressively anti-e-cigarette. (This is relatively new. Previously he focused on anti-smoking. It is not entirely clear why he has adopted this issue.) The mere fact he posted that is interesting, and he gets props for it in my mind. He did not do what most corporate tobacco control industry actors routinely do, which is to just censor comments without admitting they are doing so, with no attempt to justify their actions. Instead, Glantz basically acknowledged that he was receiving such testimonials and felt that he owed people an explanation for why he would not post them. This also contrasts starkly with the behavior of someone Glantz is often inappropriately lumped in with, the self-important pre-adolescent bully boy Simon Chapman, who has a history of aggressively ridiculing people’s personal testimonials and taking joy making fun of difficulties people have suffered. Glantz seems to be genuinely interested in being a decent person, however misguided we might think his goals.
Frankly, I have to sympathize with Glantz on this — I would not want the comments section of one of my blogs flooded with 1000 word essays that I did not consider to be informative. Ah, but there is the problem. Glantz’s claim that these posts are not informative brings us to his profound lack of understanding of the science he claims expertise in. The post in question was just over 200 words, most of which were just his unsupported personal opinions about e-cigarette regulation. The entire content that is fascinating me is:
There is a long and well-developed literature about the placebo effect where people think a treatment worked when, in fact, it was no better than a sugar pill. Personal testimonials about the benefits of e-cigarettes do not constitute scientific evidence that they are effective ways to quit smoking. If and when there are high quality longitudinal studies showing that e-cigarettes as actually used actually help people quit smoking conventional cigarettes, I will modify my opinions on e-cigarettes as cessation aids. As of now, the reality is that such studies simply do not exist.
Let’s think about the “do not constitute scientific evidence” and “longitudinal studies” bits. What exactly does he think a longitudinal study would show better than the testimonials? Such a study would consist of gathering a group of current smokers and following them over time to see what they did. Assuming these were Americans or others with similar economic and social circumstances and access to THR products, it is safe to predict that the result would be many still smoking, many becoming tobacco/nicotine abstinent, probably many switching to e-cigarettes, a fair number switching to smokeless tobacco, and a fair number switching to other nicotine products. So assuming many switched to e-cigarettes, would that constitute particularly good evidence that e-cigarettes are an effective way to quit smoking? Not really.
It would be possible that those who switched to e-cigarettes would have become abstinent (or would have switched to another low-risk alternative) had e-cigarettes not been available, and so e-cigarettes caused no smoking cessation. That would be an absurd story, knowing what we know, of course. But if we pretend, as Glantz apparently would like us to, that the study alone provides the useful evidence and nothing else does, then we would not be able to rule that out.
Why do we know that is absurd? Because of the countless testimonials by people who tried to become abstinent before they tried e-cigarettes and did not, but that quit smoking using e-cigarettes, and moreover are delighted to be using them instead of smoking. Those testimonials are the real scientific evidence, not the vague statistics that Glantz proposes collecting. Any real scientist would recognize this. It turns out that the statistics he wants to collect offer us little or no information about what we want to know — what switchers would have done without e-cigarettes. It is that kind of information that is not scientifically useful. The testimonials, on the other hand, are a rich source of scientific information about smokers who did not quit (they would probably say “could not quit”) until they found e-cigarettes. These testimonials represent useful scientific experiments.
This leads to that light switch remark in the title: I wonder what Glantz does when he walks into a new classroom [does he even teach? not sure. a scary thought for another day] and notices it is dark and not getting any lighter? Most of us would flip the switch on the wall, and if the lights came on at that moment we would conclude that that switch causes the lights to come on and be comfortable with that knowledge and take advantage of it. Glantz apparently would stand in the dark until someone conducted a longitudinal study, observing a group of rooms and what happened to the lights. But after doing that study, he would still have a problem, since the data collection would inevitably not be sufficient to discriminate between whether the switches made the lights go on, or the presence of people, or certain hours of the day. If forced to rely on that results of that study, he would still not know if that switch caused light.
Non-scientists who are being trained to appreciate the simplest forms of scientific reasoning without really learning how to think scientifically (a characterization which applies most public health and medical students) are taught that particular formal study types give us more information than do collections of less organized observations. But real scientists learn to recognize when the simplistic rules of thumb are misleading. We figure out that sometimes a collection of one-off experiments is worth far more than a formal study. Flipping a switch, or collecting a few hundred personal testimonials, is sometimes better evidence about what we want to know than any other practical study could be.
[This post is already running long, and I will probably come back to this theme later. But for those who are interested right now, I wrote a paper whose major theme is further exposition of this point for a different but remarkably related topic, which you can find here.]
I will give Glantz some more credit here, though. It is true that it is difficult to assess exactly how often something happens based on personal testimonials alone. Unfortunately, his longitudinal study would not provide very useful information about that either, since it would be for a particular population at a particular time, and would be almost impossible to extrapolate to anyone else given how rapidly circumstances are changing with respect to e-cigarettes. He also gets credit for not calling for clinical trials. Again, people who half understand science think that clinical trials are always useful, or even always best. But while they are great for studying people’s biology under fairly simple circumstances (e.g., for assessing most disease treatment options), they are generally quite poor for studying anything else, like behavior. Something like smoking cessation involves the effects of countless complicated real-world factors that are absent from an artificial clinical setting. (Of course, it is possible that Glantz’s real motive is that a longitudinal study would take much longer than clinical trials, and he just wants to stall. If that turns out to be the case, I will take back several of the things I have given him credit for.)
Having spent >1000 words dissecting his 100 words, I have still not gotten to his equally fascinating lack of understanding about the placebo effect. But discussing that will take another 1000 words, so I will leave it for tomorrow.