Skip to content

The ethics of consulting for the tobacco industry

Don Rubin published an article in 2002 on “The ethics of consulting for the tobacco industry.” Here’s the article, and here’s the abstract:

This article describes how and why I [Rubin] became involved in consulting for the tobacco industry. I briefly discuss the four relatively distinct statistical topics that were the primary focus of my work, all of which have been central to my published academic research for over three decades: missing data; causal inference; adjustment for covariates in observational studies; and meta-analysis. To me [Rubin], it is entirely appropriate to present the application of this academic work in a legal setting.

My thoughts:

I respect what Don is saying here—I don’t think he’d do this sort of consulting without thinking it through. At the same time, I think there are a couple of complications not mentioned in his article.

1. Don writes, “When I was Žfirst contacted by a tobacco lawyer, I was very reluctant to consult for them, for the standard ‘politically correct’ reasons…” I think this is a bit glib. “Political correctness” refers to attempts to restrict speech or ideology that is deemed offensive. Tobacco companies, on the other hand, actually make cigarettes, which actually do give people cancer. Now, I’m not saying that it’s immoral to work for tobacco companies, or to supply cigarettes to people who want them, or even that it’s immoral to advertise cigarettes or whatever—but to dismiss this as “political correctness” minimizes the issues here, I think.

2. Later in the article, Don presents the ethical dilemma as whether to give testimony that is scientifically valid but supports cigarette companies. In his article, he makes a convincing case that, in his analysis, the facts did not support the claims made by the anti-tobacco lawsuits.

I would tend to accept Don’s reasoning that, once he has studied the issue, it is ethical for him to call the science as he sees it, even if that means he is supporting tobacco companies in a lawsuit. (If I had close personal experience with lung-cancer victims—or with tobacco farmers—this would probably affect my views on this, but that’s another story.) However, there’s another decision point that Don didn’t spend much time on, which is his decision to work on the problem at all.

Setting aside any questions about the morality of working on the tobacco case, there is still the “opportunity cost” argument: what would have Don done if he had not worked so hard for years on this problem? Perhaps he could have made further strides in the theory of statistical modeling and causal inference, or perhaps he could have been working on an application with direct benefit (for example, collaborating with psychologists or drug designers on improved therapies or treatments). Given his involvement in the case, it is appropriate that Don did his best job as a scientist, but this still raises the question of whether he should have been involved at all.

Just to be clear: I don’t think that Don was immoral in working on this problem. In one of his books, Bill James said, “I’m not a public utlity” or something like that, and, similarly, Don should have the freedom to work on problems as he sees fit. I am not at all criticizing his ethical choices. I’m just commenting on his published article on the ethical choices. My impression from talking with Don is that he did make some progress on causal inference in the context of the tobacco study, and that one reason he worked on the topic is that it gave him the opportunity to think seriously about these problems. As I once noted, the most advanced statistical methods are often used in low-stakes problems and so it is good to see some of the most modern methods of causal inference used in this high-stakes dispute.

In any case, this article would be a great discussion-starter for a course on statistics in public health or social science. Ethical discussions in statistics can get into ruts (for example, questions of the morality of randomized clinical trials), and this article looks at a slightly different ethical dilemma that can face statisticians.

I’m curious what Chad would think of Rubin’s article. (Here’s my earlier discussion of Chad’s work on ethics and statistics.)


  1. derrida derider says:

    "I would tend to accept Don's reasoning that, once he has studied the issue, it is ethical for him to call the science as he sees it, even if that means he is supporting tobacco companies in a lawsuit."

    It is not merely ethical, but an ethical imperative. Lies in a good cause are still lies – this would be true even if you did have close experience with lung cancer victims. Just be very careful that "as you see it" is not a result of either tobacco company money or pity for victims clouding your vision.

  2. Deb says:

    "I would tend to accept Don's reasoning that, once he has studied the issue, it is ethical for him to call the science as he sees it, even if that means he is supporting tobacco companies in a lawsuit."

    The question of whether Donald Rubin's consulting for the tobacco industry is unethical depends on whether he distorted/misapplied statistics in his testimony.

    Given that:

    a. he is paid $1250/hour for consulting and $1600/hour for testifying
    b. it is notoriously easy to "lie with statistics," and
    c. it is easy to obtain transcripts of his testimony that would enable one to assess whether he lied with statistics

    shouldn't a Bayesian statistician interested in the ethical question look at the data instead of concluding "I don't think that Don was immoral in working on this problem" because "I don't think he'd do this sort of consulting without thinking it through"?

  3. Andrew says:


    I agree with you with regard to how I would proceed, or how I think I would proceed, in such a case. However, without any close experience myself with lung cancer victims, I wouldn't feel comfortable judging the perspectives of those who do.


    Thanks for the links to Rubin's testimony. I agree with you that a full study of the data could be informative here. I was making my statements about ethics conditional on Rubin's statistical reasoning. Here I am just following up on Don's article, which likewise focuses on ethical questions under the assumption that the statistical analyses were done correctly. These are interesting ethical issues in their own right, without the need to evaluate the technical claims.

    Or, to put it another way, step (c) of your reasoning would not be so easy, and would require a lot of effort that I'm not ready to spend (partly for reasons discussed in my posting above).

  4. Deb says:

    AG: Or, to put it another way, step (c) of your reasoning would not be so easy, and would require a lot of effort that I'm not ready to spend.

    Fair enough. But given that the gist of your post is "I don't have time to look at the data, but my best guess is that Rubin's testimony for tobacco companies is ethical," don't you think you ought to at least MENTION that Rubin was your thesis advisor?….

  5. Andrew says:

    Yes, Don's my mentor, collaborator, and friend. I'm so used to this that I didn't think of mentioning it above. I was assuming that my comments would be taken in that spirit. I wasn't trying to place myself as an independent judge of Don's ethical standards, but rather to explore further some of the ethical issues he raised in his article, issues that do not always come up in the usual discussions of ethics and statistics.

  6. Martin Ternouth says:

    I spent some fifteen years installing and working with large health statistics systems in the UK: at hospital, Regional and National level. In no especial order, I offer the following – which were representative of the state of knowledge and practice in the UK some ten years ago. They differ considerably from the political message.

    1. One in 200 smokers dies of lung cancer.
    2. If everyone smoked, lung cancer would be considered an hereditary disease.
    3. Someone who takes up smoking for the first time at the age of 60 will increase their life expectancy.
    4. Someone dying of a 'smoking-related disease' will be classified as such – even if they have never smoked.
    5. There is no statistical difference in life expectancy between someone who has never smoked and someone who has not smoked for five years.
    6. Smoking five or fewer cigarettes a day has no effect of life expectancy.
    7. Any detectable effects of passive smoking require a concentration of somewhere between 150 and 200 cigartettes smoked daily in an unventilated small office.
    8. Health statistics have historically in the UK been very unreliable. (In one study of an hospital group, some 50% of the patient records gave no diagnosis at all. In a study of one UK health Region, fewer than 20% of the diagnoses were permissible under ICD rules.)

    I am a non-smoker. None of my family smoke. My father gave up in 1953. I do not work for, hold no shares in, do not benefit in any way from the activities of tobacco companies. I believe that tobacco smoking is a contributory factor in ill-health and death. So why should this concern me?

    The first and most obvious effect of the demonisation of tobacco is the corresponding improvement in the health reputation of its competitors: in particular cannabis and cocaine. In the UK, where cannabis is generally smoked in spliffs of high-grade weed mixed wth 60-80% tobacco, we have the bizarre situation where one cigarette a day is considered life-threatening, but a dozen spliffs a day are 'harmless.'

    Secondly, it is apparent that we do not understand the causality. The effects of the carcinogenic agent in tobacco appear to be different – for example – to the effect of asbestos. A physician explained to me that whereas a single inhalation of asbestos dust can remain to irritate the lungs for ever, the effect of smoking is in effect to roll a set of dice with every cigarette: once you stop smoking, you start to stop rolling.

    Thirdly, smoking (or an agent in cigarette smoke – possible nicotine) – appears to have a significant effect in the delay of Alzheimer's and Parkinson's diseases. This appears to be the explanation of my fact 3 above.

    Lastly, and controversially . . .

    I was involved in one large health study (not involving smoking or tobacco) that was funded from the public purse and for which the final tranche of funding was withheld because the findings did not support the political message that the study had been commissioned to reinforce. Much as I deplore smoking, I believe that an objective study of its effects cannot now be carried out: whether funded by tobacco companies, or by the public purse. The last bastion of truth in this state of affairs is the integrity of the statistician.

  7. Andrew says:


    Are you sure that 1 in 200 smokers dies of lung cancer? I thought the risk was over 10%. For example, in 1998 there were 160,000 deaths from lung cancer out of 2,337,000 deaths total. That's 6.8% of all deaths. Considering that only 24% of Americans in 1998 smoked, the rate among smokers is gonna be much higher than 6.8%.

  8. paulse says:

    In a recent interview with Kurt Vonnegut, he said that he was suing the tobacco companies for false advertising, as he's a chain smoker and over 80 years old now.

    State governments with budget problems might think about launching campaigns against motorcycle manufacturers for not warning people about their risks.

  9. Martin Ternouth says:

    Andrew –

    I first ought to say that my knowledge of statistics is limited to a one-year module in a degree course and another in my professional studies – so I am out of my league in this company. However, I do claim some expertise in the analysis and synthesis of patterns of data.

    The figure of 1 in 200 is at least ten years old and may be different now. I should also have stated that the research that I saw and which was discussed with me related to deaths caused by smoking. More strictly therefore I should have said that only 1 smoker in 200 dies of lung cancer caused by smoking. The proposition that the research claimed to demonstrate was that lung cancer (like breast and ovarian cancer) is largely an hereditary disease. Smoking is a cause of lung cancer, but not all smokers who die of lung cancer contracted it because they smoked. Also, lung cancer may be recorded as a cause of death when it may be merely a secondary complication.

    My point here is not to defend the product of the tobacco companies but to illustrate how statistics are misused for political purposes. Three other examples from other fields:

    1. In the UK at present we have a growing problem with deaths in hospital from MRSA. The official figures are however considerably understated because MRSA is only given as a cause of death as a last resort. If the patient has a serious illness upon admission, then that is often given as the cause of death rather than MRSA.

    2. It is widely accepted in the UK that we have several hundred fatalities a year caused by 'drunken drivers'. The legal definition however refers to 'drink-related accidents'. Of these, many – a majority in some years – are accidents where the driver is sober but a pedestrian is drunk. Others relate to accidents where the drunken driver was parked or otherwise stationary, or where the other (sober) driver was on the wrong side of the road. In one recent year – after making these adjustments – we had the odd anomaly that sober drivers caused more deaths in accidents where one of the drivers was drunk. Again, my purpose is not to condone drunken driving but to illustrate a political misuse of data – which is the substance of the discussion in this thread.

    3. The third example is anecdotal and relates to the incidence of New-variant CJD in humans. There is some (but limited and disputed) evidence that the disease may be transmitted by eating meat from animals infected by BSE. There is difficulty in establishing the exact cause because of the minute number of cases. The BSE-ingestion hypothesis is appearing more doubtful as time passes because deaths from N-vCJD on this hypothesis should now be well over 100,000 in the UK but have stayed resolutely in the low dozens. A few years ago a young woman died of the disease. She and her parents had been life-long vegans. The parents, looking for some explanation in their distress, fixed upon the idea that their daughter must at some time eaten infected meat without their knowledge – perhaps at a birthday party or a school outing. This story duly surfaced in the UK national press under the headline 'JUST ONE HAMBURGER CAN KILL!'

  10. Deb says:

    I wrote a response to this post on my blog. The trackback function is not working, though. See:

  11. Martin Ternouth says:

    An article in Nature, 9th June 2005, about how funding affects research results.

  12. Garrett Glasgow says:

    Deb's comments on how much Rubin makes per hour of consulting and Gelman's relationship to Rubin are examples of a logical fallacy known as "poisoning the well," where the goal is to discredit the source of the argument rather than address the argument itself. For instance, demonstrating that Rubin did in fact distort or misapply statistics in his testimony would be a valid criticism of his conclusions — an ad hominem attack based on how much money he makes is not.

  13. Deb says:


    I disagree that my comments re:

    a. Gelman's relationship to Rubin
    b. Rubin's remuneration and

    were examples of "poisoning the well." As a Bayesian, I think of those comments as evidence.

    AG's original post was a review of Rubin's article. Journal editors and funding agencies require that reviewers reveal their relationships to authors. At the National Science Foundation, the advisor-advisee relationship is considered so sacred, that it is NEVER kosher for an advisor to review an advisee's proposal (or vice versa), even 40 years after the Ph.D. was awarded.

    AG agreed with the claim made in DR's paper that DR's motivation for consulting was the belief that statistics was (were?) being abused by the plaintiffs. In order to evaluate the likelihood of this hypothesis, we need to know p(alternative hypotheses). The fact that DR was paid millions for consulting increases the likelihood that he had non-intellectual motives for consulting.

  14. Garrett Glasgow says:

    My apologies — I think I misinterpreted your earlier comment made on October 11. If you were commenting on the *statistical conclusions* from Rubin's work, pointing out how much money he made would have been "poisoning the well" — hinting that the results of his testimony were bought with tobacco money rather than arrived at honestly. In that case my original criticism would stand — critics should point out the statistical flaws in the study rather than engage in ad hominem attacks. However, if commenting on his *motives* for undertaking this consulting work I don't see any problem with treating his compensation as data.

  15. Barry says:

    I'd add that the tobacco industry seems to have been the modern fountain of the junk science for PR industry. Working for them would be dangerous, IMHO, like working for organized crime – once you get involved, don't expect to become uninvolved.