Skip to content

“If our product is harmful . . . we’ll stop making it.”

After our discussion of the sad case of Darrell Huff, the celebrated “How to Lie with Statistics” guy who had a lucrative side career disparaging the link between smoking and cancer, I was motivated to follow John Mashey’s recommendation and read the book, Golden Holocaust: Origins of the Cigarette Catastrophe and the Case for Abolition, by historian Robert Proctor.

My first stop upon receiving the book was the index, in particular the entry for Rubin, Donald B. I followed the reference to pages 440-442 and found the description of Don’s activities to be accurate, neither diminished nor overstated, to the best of my knowledge.

Rubin is the second-most-famous statistician to have been paid by the cigarette industry, but several other big and small names have been on the payroll at one time or another. Here’s a partial list. Just including the people I know or have heard of:

Herbert Solomon, Stanford
Richard Tweedie, Bond U
Arnold Zellner, U of Chicago
Paul Switzer, Stanford
Joseph Fleiss, Columbia
Nathan Mantel, George Washington U
Joseph Berkson, Mayo Clinic

Also the well-known psychologist Stanley Schacter, sociologist Peter Berger, and Ernest Hook, who’s not so famous but whom I happened to know because he sometimes would hang out at the faculty lounge at the UC Berkeley statistics department. Of all these people, the names that surprise me the most are the public health researchers such as Fleiss.

Much of the cancer-denial work was done after the 1964 Surgeon General’s report. For example,

The statistician George L. Saiger from Columbia University received [Council for Tobacco Research] Special Project funds “to seek to reduce the correlation of smoking and diseases by introduction of additional variables”; he also was paid $10,873 in 1966 to testify before Congress, denying the cigarette-cancer link.

And here’s a famous name:

Ingram Olkin, chairman of Stanford’s Department of Statistics, received $12,000 to do a similar job (SP-82) on the Framingham Heart Study . . . Lorillard’s chief of research okayed Olkin’s contract, commenting that he was to be funded using “considerations other than practical scientific merit.”

Ouch. I bet that one didn’t make it into the Stanford alumni magazine.

As late as 1974, a cigarette-company-funded pharmacologist “published an article in Executive Health titled “The Case against Tobacco Is Not Closed: Why Smoking May Not Be ‘Dangerous to Your Health’!”

This does not fit well with cigarette lobbyists’ claims that everybody knew all along that cigarettes are dangerous, people used to call them “cancer sticks” etc etc. As Proctor demonstrates, surveys over the decades have found a lot of uncertainty about the health risks of cigarettes—and the cigarette companies were doing their best to prolong this uncertainty.

One thing I learned from Proctor’s book was the distinction between tobacco and cigarettes. Tobacco is bad for you, cigarettes kill. What’s the difference? Two biggies: mass production and how the tobacco is processed. Mass production means that higher doses are more convenient and affordable (not such a good thing if you’re addicted to a product that causes cancer). The part I didn’t know about, before reading this book, is that the physical/chemical treatment (in particular, something called “flue-curing”) makes cigarettes much less irritating to the throat, so that a smoker can more easily inhale and get those carcinogens directly into the lungs.

Thus, a world in which people grew tobacco in their backyards and rolled their own cigars would cut out lots and lots of smoking morbidity and mortality.

One thing I didn’t quite catch, though—I’ve never puffed on a cigarette myself—is my the nicotine patch isn’t more popular. If people want to quit, why not give the patch a try?

Proctor also notes the divergent interests of two groups that are often conflated, “cigarette companies” and “smokers.” According to surveys cited by Proctor, most smokers (in the U.S., at least) want to quit. I’m assuming that most cigarette companies don’t want this. This is all well known but it sometimes gets lost in discussions of the “war on smokers,” etc.

Some juicy bits from Proctor’s book:

Proctor reports the following amazing court testimony from “Kenneth Ludmerer, M.D., a Washington University professor of medicine and medical history. . . . Ludmerer says he agreed to work for the industry after seeing the poor quality of historical testimony introduced by the plaintiffs, but what is remarkable is how truncated his own investigations have been . . .”:

Ludmerer made this statement in 2002. I looked him up and he still seems to be employed.

I wonder if he still has no opinion on whether “cigarette smoking contributes to the development of lung cancer in human beings”? Maybe there have been some important research developments since 2002 that have convinced him. I bet it’s tough for the poor guy, being such a lonely voice in the wilderness. At some level, ya gotta admire someone who’s willing to make a statement that will make him appear to outsiders as either a liar or a fool. Ludmerer reportedly received over half a million dollars for his testimony, but I can’t believe he needed the money. Cost of living in St. Louis is, like, nothing, and I have a feeling Wash U. already pays him pretty well. I can only conclude that he either believed everything he said or that he felt there was a principled reason to lie about his beliefs.

The big picture

As a professor, it is natural for me to get particularly indignant about the offenses of my academic colleagues. To his credit, though, Proctor doesn’t lose track of the larger story, which is the century-long transition of tobacco-smoking to mass-produced lethality. In the first part of the twentieth century, cigarette companies conducted research into smoking and cancer, with the hope of developing a safe cigarette. After all, they had no motivation to kill their customers. It eventually became clear that the safe cigarette wasn’t going to happen.

Here’s Proctor:

This all makes a lot of sense, and it’s a lot different from the picture I got from talking with Rubin about this, several years ago. Then the story was that everyone had known forever that smoking caused cancer, and that cigarette manufacturers were performing the useful service of supplying a consumer good that many people wanted. It’s interesting to see that, at least in public, cigarette executives taking a much more direct position that they did not want to be in the position of giving people cancer: “If our product is harmful . . . we’ll stop making it.”

P.S. I wrote this awhile ago but it kept getting bumped. In the meantime, I used much of it in this ethics column for Chance.


  1. Great post and great Chance article. From the latter “The rules promulgated by Austin Bradford Hill for assessing causation in observational
    studies—inspired in large part by studies of smoking and cancer—remain influential and have recently been re-expressed and extended by Paul Rosenbaum into a statistical framework of causal inference”. Can you give references for those things?

    Also, I very much hope you will do a Ethics column on defense / military funding; you touch on it in the above-mentioned but it really deserves a whole article. It is also a very real issue for many of my colleagues.

    • Brian says:

      The Hill criteria ref is:

      HILL, A. B. (1965). The environment and disease: Association or causation? Proc. Roy. Soc. Med. 58, 295–300.

      They’re generally used as rules of thumb for causal inference, rather than being part of a formal framework (in fact Hill was pretty clear that this was how they should be viewed). There’s the occassional attempt to express them algorithmically, but they rarely get much traction. In fact causal inference in public policy – at least at the EPA – has if anything shifted towards a more qualitative approach in the past few decades.

  2. A. Zarkov says:

    “Rubin is the second-most-famous statistician to have been paid by the cigarette industry, but several other big and small names have been on the payroll at one time or another.”

    How about R. A. Fisher? He was a consultant to The Tobacco Manufacturers Standing Committee. Both Fisher and Berkson (they agreed on little) shared criticisms about the case against tobacco. Berkson was both a statistician and a doctor as well as the head of the statistics group at the Mayo Clinic. Mayo had, and has, a long-term large database for him to use. As such he was well positioned to have an informed opinion on the controversy. I don’t know if their objections have ever been fully answered. B.W. Brown at Stanford University School of Medicine has a chapter in Statistics: A Guide to the Unknown, Tanur and Mosteller eds. cover Fisher, Berkson and their role in the controversy. In y opinion, well worth reading to get a more dispassionate picture.

    Andrew, your whole post gives the impression that some or most statisticians who advised, consulted or worked for the tobacco industry must have been corrupted by them. I really doubt that applies to men like Fisher and Berkson. They had valid objections based on the state of knowledge at the time.

    I once worked with a doctor who had lung cancer, and had a lung removed. Yet he remained a chain smoker. I remember he smoked continuously during meetings. My own doctor had the same thing happen. I remember having lunch with a famous statistician who wrote a well-known book on multivariate statistics. She smoked during the lunch, and told “the risk of small. BTW what is the risk? How does that compare to cannabis? In the counter culture, which seems to rule California today, smoking tobacco is verboten, but Marijuana is fine.

    • Andrew says:


      1. Yes, when I referred to Rubin as the second-most-famous, I was thinking of Fisher as the #1 guy.

      2. Just a suggestion: Next time you want to make the argument that cigarette smoking might not cause lung cancer, I suggest you don’t use an anecdote about a chain smoker who had lung cancer!

      3. Yes, I think it’s agreed that smoking the occasional cigarette is not such a risk, especially if you don’t inhale. It’s that hard-to-shake lifetime habit that’s the real public health problem.

      4. My impression is that it’s considered rude and illegal in many places to smoke tobacco or marijuana indoors, but that people don’t have much of a problem with either of these if you’re outdoors and don’t blow the smoke in other people’s faces. But maybe California is different—do they smoke marijuana in the office there? That would indeed be rude.

  3. Anatoly Karp says:

    Andrew, did you examine the existing evidence for the smoking-cancer link? It might be a good a idea to apply your skeptical powers to the subject. Who knows, maybe you’ll become a cancer-denier yet.

  4. A. Zarkov says:


    1.So why didn’t you come out and refer to Fisher? I don’t think everyone knows that Fisher was critical of the tobacco-cancer link.

    2. I guess I didn’t express myself well. I am not making an argument “that cigarette smoking might not cause lung cancer.” Quite the contrary. The anecdotes I cited show that some people, no matter how well informed as to the risk, still choose to smoke. They make a trade-off between enjoyment and risk. I personally would not make that trade-off (I have never smoked even a single cigarette), but I understand and respect those who do. I am fully convinced that cigarette smoking causes lung cancer, and other serious degenerative diseases. I would never encourage anyone to smoke. Indeed I convinced my mother to stop decades ago, and she did.

    3. While we have abundant and compelling evidence today that cigarette smoking causes cancer, this was not always the case. Fisher and Berkson got it wrong. But they got it wrong for good reasons. The case for secondary smoke is another matter.

    4. Many places do ban smoking– even indoors– even in your own home or apartment. Some jurisdictions ban smoking at the beach! New York City (where I’m from) is headed in this direction. In California it’s socially declasse to smoke tobacco anywhere. Indoors, outdoors, in your car, on a bicycle. Smoking in offices is banned virtually everywhere as well as restaurants, bars, clubs etc. Yet the very same people who scoff at those who smoke, are tolerant of those who smoke Marijuana as though it were absolutely harmless. I think they’re being hypocritical.

    • Andrew says:


      1. Sometimes I try to makes things interesting by not spelling everything out.

      2. Addiction is not always the same as enjoyment.

      3. Yes, much has been learned since the 1950s.

      4. I don’t see anyone smoking marijuana indoors, nor do I see them smoking marijuana in places where I don’t also see people smoking cigarettes. I don’t recall ever seen anybody smoking marijuana on a bicycle either! But I don’t live in California; maybe things are indeed different there.

  5. A. Zarkov says:

    In California go to an indoor rock concert and you will smell lots of marijuana. Or a night club, many private parties. Perhaps New York City is different. It wasn’t when I lived there, but that was 30 years ago. I was kidding about bicycles. Sometimes I try to make things interesting.

  6. HowardC says:

    * “Andrew, did you examine the existing evidence for the smoking-cancer link? It might be a good a idea to apply your skeptical powers to the subject…” A.K.


    Yes, yes — what exactly does the current statistical evidence prove, beyond a reasonable doubt ?

    Like A.Z. here, I do not doubt the causal smoking-cancer link, but I must admit I’ve never actually seen conclusive evidence of it published publicly. It must certainly be out there somewhere, on this very critical health issue.

    What are the facts ?

    • See Sec. 8 of Freedman’s paper “From association to causation: Some remarks on the history of statistics.” Statistical Science, vol. 14 (1999) pp. 243–58, Reprinted in Journal de la Société Francaise de Statistique, vol. 140 (1999) pp. 5–32 and in Stochastic Musings: Perspectives from the Pioneers of the Late 20th Century, Lawrence Erlbaum Associates (2003) pp. 45–71, J. Panaretos, ed., available at , for a 4-page discussion of the way that epidemiologists concluded that smoking causes fatal diseases, with references.

    • Andrew says:

      Also, the Wikipedia page looks pretty comprehensive.

  7. John Mashey says:

    1) Fisher was mentioned in comments in earlier discussion. As I noted, search “ronald fisher” in the tobacco archives. Whether or not the evidence was enough, it is clear that he looked for every reason for it not to be. See Proctor, p.436 on ‘simply “for the money.”‘

    2) Nicotine is highly addictive, especially if a smoking habit starts early enough, like 12-18, which is why cigarette companies have long focused their marketing on kids, since very few people start after 18. Read Proctor’s book, or see PDF here, pp.37-38, especially “The Importance of Younger Adults” and Figure F.3, and the terrific success of Joe Camel. Cigarette smoking is really not an adult choice.

    3)Not being any kind of smoker, I don’t know much about cannabis addiction, but I find Is Marijuana Addictive?, who answer seems to be: yes, for some, like alcohol, but nowhere near like tobacco, especially that in cigarettes. (Andrew noted the key discussion from Proctor that cigarettes != tobacco, but are designed to be as addictive as possible.)

    4) Hence:
    cigarettes: strongly addictive, addiction acquired by age 18, most adults would rather stop if they could, secondhand smoke long shown to be damaging, especially to children. Marketing efforts by the “best” marketeers in the world, also great lobbyists.
    I’ve know lots of smart, determined people who simply could not stop.

    marijuana: much less addictive, people find it easier to stop if they want to, much more of an adult choice (like alcohol). I don’t know about secondhand smoke, but I suspect that indoors pot-smoking mostly sticks to venues and events where people might expect it. Of course, for some people there are real medical benefits. Some people who drink wine will turn into alcoholics, but CA is not likely to ban wine, since most wine-drinkers drink quite moderately, and have not the slightest wish to stop.

    Given important objective characteristics that differ, it seems not at all hypocritical here in CA to try to stop the first while allowing more room for the second. Personally, I think the less junk in the air the better, and it is way better than when I was dating a woman with asthma problems, where we had to be careful where we went in NJ or NYC.

    • A. Zarkov says:

      The NIH paints a somewhat different picture with regard to marijuana than the pop psychology magazine Psychology Today. While they don’t report a link between marijuana use and lung cancer, they do say,

      “Numerous studies have shown marijuana smoke to contain carcinogens and to be an irritant to the lungs. In fact, marijuana smoke contains 50-70 percent more carcinogenic hydrocarbons than tobacco smoke.”

      Moreover cigarette smoking does not impair mental functioning. One can smoke cigarettes and play chess, prove theorems, write novels, and drive a car. Who would you be more comfortable with, a driver who was smoking cigarettes or marijuana? Indeed the NIH says,

      “Not surprisingly, marijuana intoxication can cause distorted perceptions, impaired coordination, difficulty with thinking and problem solving, and problems with learning and memory. Research has shown that, in chronic users, marijuana’s adverse impact on learning and memory can last for days or weeks after the acute effects of the drug wear off.2 As a result, someone who smokes marijuana every day may be functioning at a suboptimal intellectual level all of the time.”

      So I think the jihad against tobacco and tolerance of marijuana is hypocritical. Again this does not mean I think cigarettes aren’t harmful. If everyone stopped smoking, tomorrow I would be pleased. I would also be pleased if everyone stopped smoking marijuana as well. It is worse than drinking wine. Ethanol metabolizes quickly compared to cannabis.

      Finally I note that David Freedman wrote, “Since lung cancer is a rare disease even among smokers …,” I suspect the secondary smoke risk of lung cancer is extremely small because the dose from the active agents should be very low compared to smokers. Of course it all depends on how much you smoke and how long you get exposed to secondary smoke. The could also be other respiratory effects that would justify regulation of secondary smoke sources.

      See for the NIH material.

      • revo11 says:

        From what I remember the PNAS+ paper relating marijuana and long term learning didn’t come anywhere near establishing causality. The only thing that came out significant was the cohort that started smoking at a very young age and there’s a lot of obvious confounders associated with individuals that smoke marijuana before they’re 14 that are nowhere in their analysis.

        Cigarette smoking habits and marijuana smoking habits are very different. It’s common for smokers to smoke a pack or more a day. I suspect few people smoke marijuana that frequently for as an extended period of life as tobacco smokers. “Moreover cigarette smoking does not impair mental functioning” – why would you just focus on “mental functioning” as a measure of interest? If you pick and choose harmful side effects you can make anything less harmful than anything. Also, the fact that the marginal frequency of lung cancer within smokers is small doesn’t mean that there could be larger heterogeneous effects within the population (for example, conditional on consumption). I haven’t delved into this literature, but I doubt “smokers” represents an exchangeable population.

        Anyway, to say there’s a “jihad against tobacco and tolerance of marijuana” is absolutely disingenuous. How many politicians in power in favor have spoken out in favor of legalizing marijuana? How many politicians in power have spoken out in favor of making tobacco illegal?

  8. John Mashey says:

    If people have never seen conclusive evidence, they may want to actually look for it.

    Proctor spends pp.145-252 on the history of this, including lots of pointers to documents where the cigarette vendors themselves admitted the link. It’s easier to read that than paw through the tobacco archives, but the latter can be quite revealing.

    If the cigarette vendors knew they caused cancer, one hardly needs to look at the medical studies, but if one wants to, that’s easy.
    See CDC on tobacco, whose website has:

    a)summarized conclusions

    b) Detailed Surgeon General reports, such as this one, whose Chapter 1 has statistical commentary. Then, the 326p Chapter 2 goes through the different forms of cancer, presents the evidence about each, with 30+ pages of references to medical studies.
    Other chapters cover other forms of diseases, and all these calibrate the strength of evidence.

    But again, all on really needs to know is that the cigarette companies knew cigarettes caused cancer by the early 1950s. For instance, read about Ecusta (pp.210-223) in Proctor’s book.

  9. […] blog has a largely overlapping post, but with a bit more material (and discussion in […]

  10. Jonathan (a different one) says:

    As an expert witness myself,I think you confuse (somewhat) the role of an expert. heexpert is there to assist the Court on things that are in the scope of his testimony. They are not there to tell the Court how to decide the ultimate issue. (Didn’t the lawyers tell you this when you testified?) You may have lots of knowledge which is beyond the scope of your expert statement, and it is almost certainly in the interests of the side hiring you to narrowly circumscribe what your testimony is about (there are some exceptions to this, but not many). Thus, they will pay you to examine X, but not X’. If it is impossible to have a coherent opinion about X without studying X’, you tell the lawyers and they either fire you or let you study X’. But an opinion about whether or not smoking is harmful or not or what PM felt about it may well have had nothing to do with his testimony. The internal documents themselves were presumably already on the record already and unless his opinion on those hep the Court, what’s the point?

    • Andrew says:


      I’ve never actually testified in court but, yes, it seems that Ludmerer was doing what he was paid to do. The sad thing is that he was paid to say that he had “no opinion” on whether cigarette smoking contributes to the development of lung cancer in human beings. No opinion, huh? That’s just pitiful. I don’t know how much this guy was being paid, but you couldn’t pay me enough to say something like that.

      • Jonathan (a different one) says:

        But an expert opinion would have to have a basis… and it couldn’t just be anecdotal, or the opinion of a normally well-educated person. If he didn’ study it, it’s obviously the right answer. To take an example, I have testified about statistical adverse impact in a reduction in force at a corporation. If the opposing attorney asked me whether or not I thought there is some generic level of age discrimination in America my answer would generally be: “Any opinion I have would be a layman’s opinion, which has no value to the Court at all. I’m testifying about certain calculations I made in this case; calculations that the Court lacks the expertise to properly formulate and make. But my opinion about age discrimination is not part of my testimony and is nothing I’ve studied.”

      • A. Zarkov says:

        They don’t do science in court– it’s mainly theater. Most juries don’t understand the expert witness at all. If he seems credible by his demeanor, dress etc, they will get duly influenced. Judges don’t understand anything either. Many of them barely understand the law, let alone physics and statistics. The better ones get a special master to help them understand what’s going on with regard to the technical matters.

        The evidence behind the breast implant suit against Dow Corning was bogus. My daughter studied that case in law school. I asked her if the plaintiffs had to give the money back after the evidence came apart. She responded, “funny thing dad, I asked the law professor the same question, the answer was “no,” they don’t have to give the money back.”

        Remember U.S. and British law is an adversarial process. It’s a contest with strict rules. As an expert witness you don’t get to pontificate or give your personal opinion. You have to stick to the narrow issue at hand, or the judge will shut you up. The adversarial process is supposed to lead to truth. Some people have said it’s like trying to use two blurry photographs to get a clear one. In complex cases it rarely does.

  11. John Mashey says:

    The point of the expert witness tactic is to get people who can raise doubts on very narrow grounds, but always have an excuse not to address contradictions or context. For what it’s worth, this is yet another tactic picked up from the tobacco folks by climate anti-science, of which the best example if Andre’s “favorite” statistician, Ed Wegman.
    See PDF @ Strange Scholarship in the Wegman Report, specifically Meme-h, “We’re just statisticians.” (Do a Full Search for Meme-h to see examples).

    There, the goal was to cast doubt on the 1999 hockey stick, by emphasizing the use of decentered PC there (which as wrong), without ever answering the questions:
    1) But did that error actually make any significant difference? (NO)
    2) Had that error long before been corrected and replaced by better ones? (YES)

    Of course, Congressional hearings are sometimes described as Kabuki theatre, and I suspect some legal cases may be similar.

  12. A. Zarkov says:

    McIntyre and McKitrick respond to the claim that the lack of PC centering “makes no difference.” Moreover the Mann algorithm will give hockey stick like results when driven by correlated noise.

    • John Mashey says:

      False statement, long out of date, and irrelevant to the topic of the post, which statisticians doing odd things, soemtiemns with regard to expert witnessing.

      See Replication and due diligence, Wegman style and the further Effect of selection in the Wegman Report. As Nick rote:
      ‘-the HS PC’s shown were anything but random samples. In fact, the 10000 simulations had been pre-sorted by HS index, and the top 100 selected. A choice was then made from this top 100.

      – Although Wegman had said that “We have been able to reproduce the results of McIntyre and McKitrick (2005b)”, the PC in Fig 4.1 was identical to one in MM05b. Since the noise is randomly generated, this could not have happened from a proper re-run of the code. Somehow, the graph was produced from MM05 computed results.

      – The red noise used in the program was very different to that described in the caption of Fig 4.4.’

      MM generated graphs with unrealistically-high persistence, then selected the 1% with the biggest positive hockey sticks. Wegman and co picked this up, either because they didn’t understand it, (quite possible, given the other statistical incompetencies) or didn’t care. I’ve looked at MM’s code, and Wegman still owes his version to Waxman, as per this. At least one fine statistician, Noel Cressie, asked to review the WR, told Wegman he should show the downward hockey sticks also. That never sank in.

      Now, how about back to statisticians, ethics, tobacco and/or expert witnessing.

      For Andrew and others, on ethics:
      Consider a scale from 1 to 10:
      10: Someone seriously believes the evidence is inadequate. (After all, bad studies exist.)
      Testifies, with no compensation, except possibly expenses.

      5: started at 10, but the money got good, and still thinks they are at 10, but really not.

      1: doesn’t care in the slightest about truth, in it for the $.

      The interesting question is where people might have started and how they migrated, if they did.
      I’ve studied some of the sequences around climate anti-science, but not for statisticians+tobacco.

  13. Moetner says:

    Wikipedia is the definitive source for a comprehensive smoking-Lung-Cancer causal link ?

    Vague statistical inference can not possibly establish such a causal link.

    Even valid associative inference should establish a 50-100% correlation between smoking and cancer, but it does not even come close. Most people who smoke don’t get lung cancer, and at least 10% of Americans who do get lung cancer- do not smoke. There are also huge international/ethnic variations among smokers and cancer rates.

    There is currently no proof whatsoever for the alleged smoking-cancer causal link. None.

    Smoking is a disgusting and silly habit. But all that one can now objectively say is that it is a risk factor for cancer and increases the incidence of lung cancer.

    • Andrew says:

      1. I wrote that the Wikipedia page “looks pretty comprehensive,” not that it was definitive.

      2. If a correlation is lower than 50%, it’s lower than 50%. That doesn’t make it invalid, it’s just the way the world works.

      3. Regarding your claim of no proof etc., you’ll be happy to learn (from the above blog post) that at least one professor at a leading medical university agrees with you. So maybe you and Prof. Ludmerer are right. After all, everyone laughed at Pons and Fleischmann and now here we are with cold fusion generators in every basement!

      • Joseph says:

        I agree that the association between smoking and lung cancer is not a deterministic one. But very few things about human health are. But the strength of the association is pretty strong (at least for primary use cigarette smoking) and the level of evidence is as strong as we can get without randoming people to be smokers as part of an experiment. I refer you to:

        It is not a perfect study. But I see it as above the “no proof whatsoever threshold”; consider Table 6 (for example) which seems to show an non-trivial association.

        • Brian says:

          I thought at first Moetner was trying to highlight the limits of statistical inference in general. But his comment about “valid associative inferences” is just plain weird. Probably best to move on.

  14. Sanford says:

    I am by no means an expert on the subject and have never smoked but has anyone considered the increased use of dangerous chemicals/pesticides/additives in the growth of tobacco and production of cigarettes over the last century as a contributing factor to lung disease? Is it possible that the pesticides and chemical additives are the culprits here if there is any culprit at all?

    Regarding marijuana use – please be aware of the harmful effects for adolescents who experiment even once. These young people are especially vulnerable because they are still in their developmental stages and if they have a genetic predisposition for mental health issues, one time use can trigger an irreversible state of schizophrenia. I am a retired ECE prof and was well aware of the latest research and findings on this alarming issue that was just recently coming to light over the last decade. One of my closest friends works in mental health for youth and is very saddened to be confronted with many cases in his practice. Marijuana is NOT harmless!

    • John Mashey says:

      Has anyone considered…?
      yes, see Proctor’s book, especially Section Four, but it pops up elsewhere. The tobacco companies did a lot of research hoping the problems were do only to something other than burning tobacco. No such luck, but some of the other junk surelydoes not help, either.

      For instance, for a few years, Kent “Micronite Filters” were asbestos. Great.

      Marijuana: I don’t know that turf, but I think the issue is that each problem has to be dealt with appropriately to any differences. In general, societies try to protect children against irreparable consequences of bad decisions, while giving them enough room to make mistakes that won’t damage them permanently.

  15. Klaus K says:

    Re. Smoking and lung cancer: You may take interest in the Scottish “McTear case” from 2005 against Imperial Tobacco. The case was brought on by the British anti-smoking forces (including R. Doll) on behalf of a widow whose husband died of lung cancer. After considering the published papers and hearing many witnesses on both sides through 12 years Judge Nimmo concluded:

    “The judge found that the results of animal experiments conducted in the 1950s and 1960s failed to support the general causation hypothesis. Although this was not a crucial point in itself, it was nevertheless significant, as proof of a causal connection between cigarette smoking and lung cancer would depend solely on the conclusions from epidemiological studies. The judge stated that, on one view of the scientific approach, these studies “could be regarded as yielding no more than untested hypotheses”.

    “It followed that the claim would stand or fall on the basis of what could be proved by the epidemiological studies. After hearing evidence on general causation, the judge was unable to decide whether cigarette smoking causes lung cancer. He was critical of the evidence given in this regard by the claimant’s expert [Richard Doll] and concluded that there was no case for Imperial to answer.” – subscription needed

  16. Ernest B. Hook says:

    I just stumbled upon this exchange. I suspect that “Andrew” is Andrew Gelman, who got a raw deal at Berkeley, and I am certain is doing well now at Columbia. His departure was a great loss to the Berkeley campus, and applied statistics here, although some theoretical statisticians unfortunately did not recognize that fact.
    But whether this is by Andrew Gelman or another, I can understand how Proctor’s presentation can have misled.
    First, as to the association of smoking with lung cancer and other adverse health outcomes, I strongly believe smoking causes these. Not all who smoke will develop the adverse outcome, not all with the adverse outcome will have smoked (albeit they may have been exposed to 2nd hand smoke), but a causal connection is there.
    One would have to argue otherwise that the entire association results from some factor X which “causes” i.e. predisposes, to both smoking and cancer independently. I regard this as so highly implausible, it is not worth considering, irrespective of judicial decisions cited etc. . Early on, many statisticians, some prominent, felt that because that POSSIBILITY had not been FORMALLY excluded, a PLAUSIBLE connection of smoking with cancer could NOT be accepted.
    Tobacco companies in my view make a product that causes harm. But that is entirely separate from a different issue, a historical one. Accepting that cigarette smoking is toxic, after recognition of a likely causal association, what did experts recommend early on, that tobacco companies do? That is the historical question at which I looked. The experts recommended a filter that would remove the toxins. There is an overwhelming literature indicating this. And companies tried to do that; unfortunately the filters did not work for a number of technical reasons. Nevertheless, that was the recommendation of the experts at the time, and the companies did attempt to implement that advice. There are many things I believe one can fault the companies for. But the historical record on that specific matter regarding introduction of filters shows an attempt to comply with suggestions from various “experts”.
    In his personal crusade on this issue, Proctor’s analysis and listing of people “defending” the companies glosses over the great variety of viewpoints and matters addressed. It misleads seriously, and indeed in my opinion borders on attempts at character assassination.

    • Andrew says:


      Thanks for the background. My impression from reading Proctor’s book was that he wrote the same thing you wrote above, that the cigarette companies did a lot of research into filters to remove the toxins, but then at some point they learned that filtering doesn’t really work. But they did not ever stop making cigarettes (despite the above-reproduced quotes from Weissman, Bowling, Bible, Judge, and Long, stating that they would stop if they were convinced that smoking was harmful).

  17. […] Collins (here’s a link but I don’t know the original source) picked up on my article about statistics and the cigarette companies, which mentioned a consulting job of famed Stanford […]