Blowing smoke

Who wouldn’t agree with this claim of Dr.Gio Batta Gori, of the Cato Institute (published in today’s Washington Post):

>Presumably, we are grown-up people, with a civilized sense of fair play, and dedicated to disciplined and rational discourse. We are fortunate enough to live in a free country that is respectful of individual choices and rights, including the right to honest public policies. Still, while much is voiced about the merits of forceful advocacy, not enough is said about the fundamental requisite of advancing public health with sustainable evidence, rather than by dangerous, wanton conjectures.

That admirable goal, however, is not advanced by this sort of thing:

>Lung cancer and cardiovascular diseases develop at advancing ages. Estimating the risk of those diseases posed by secondhand smoke requires knowing the sum of momentary secondhand smoke doses that nonsmokers have internalized over their lifetimes. Such lifetime summations of instant doses are obviously impossible, because concentrations of secondhand smoke in the air, individual rates of inhalation, and metabolic transformations vary from moment to moment, year after year, location to location.

I’m not a scientist. But even I can tell the difference between reasonable objections to basic methodology and pushing goalposts back a little (or in this case a lot further). For, on Dr.Gori’s argument, assessing the effects of secondhand smoke is “impossible.” He continues:

>In an effort to circumvent this capital obstacle, all secondhand smoke studies have estimated risk using a misleading marker of “lifetime exposure.” Yet, instant exposures also vary uncontrollably over time, so lifetime summations of exposure could not be, and were not, measured.

>Typically, the studies asked 60–70 year-old self-declared nonsmokers to recall how many cigarettes, cigars or pipes might have been smoked in their presence during their lifetimes, how thick the smoke might have been in the rooms, whether the windows were open, and similar vagaries. Obtained mostly during brief phone interviews, answers were then recorded as precise measures of lifetime individual exposures.

>In reality, it is impossible to summarize accurately from momentary and vague recalls, and with an absurd expectation of precision, the total exposure to secondhand smoke over more than a half-century of a person’s lifetime. No measure of cumulative lifetime secondhand smoke exposure was ever possible, so the epidemiologic studies estimated risk based not only on an improper marker of exposure, but also on exposure data that are illusory.

Don’t forget to undermine the credibility of the witness:

>Adding confusion, people with lung cancer or cardiovascular disease are prone to amplify their recall of secondhand smoke exposure. Others will fib about being nonsmokers and will contaminate the results. More than two dozen causes of lung cancer are reported in the professional literature, and over 200 for cardiovascular diseases; their likely intrusions have never been credibly measured and controlled in secondhand smoke studies. Thus, the claimed risks are doubly deceptive because of interferences that could not be calculated and corrected.

Lastly, there are good arguments on both sides:

>In addition, results are not consistently reproducible. The majority of studies do not report a statistically significant change in risk from secondhand smoke exposure, some studies show an increase in risk, and ¿ astoundingly ¿ some show a reduction of risk.

A more reasonable interpretation of that situation would be this: assessing secondhand smoke is very tricky, and much like assessing any cancer risk, it involves probabilities and factors that often elude the kind of painful exactitude we would like to demand from our science, but as a matter of fact, almost never get. No one but those ignorant of statistics and the meaning of basic scientific studies–oh, I can think of certain Chief Executives and senators–would make such an absurd demand on that kind of evidence. The bar is too high: count how many times Dr.Gori says “impossible” or denies that the thing in question is not subject to proof. That’s a very decisive conclusion for a scientist. Few I think would agree that such a thing is impossible from the outset. Difficult maybe. But not impossible.

Finally, the careful reader will also note the very narrow scope of Dr.Gori’s analysis: to deny that the evidence shows a decisive causal connection between secondhand smoke and lung cancer and cardiovascular disease. Even if he’s right about that specific and narrow claim, that doesn’t mean secondhand smoke doesn’t play a role in either increasing the risk for those ailments or in exacerbating others.

In the end, there is a more basic question of burden of proof here. At this point, the burden rests with the one who claims that living and working in a smoky environment isn’t bad for you.

11 thoughts on “Blowing smoke”

  1. Your statement of the burden of proof here is quite dubious. The last I checked, the burden of proof is on anyone making a positive causal statement. In this case, the positive causal statement would be that secondhand smoke causes health harms. Surely the burden of proof is on those who allege that this causal relationship exists. Burden of proof is determined by the nature of the claim, not by whether or not it is a claim that CEOs and senators are likely to make.

    You criticize this author for ignoring basic knowledge of statistics, yet you never contest his claim that statistical studies are inconclusive about the effects of secondhand smoke. You also never contest his claim that lifetime exposure is never actually measured by these studies. You claim that it is dubious to assert that these measurements are impossible, but in that case you owe us an explanation for why these measurements were not actually taken. If the author is correct that they weren’t, it seems that the impossibility–or at least extreme difficulty–of such measurements could be a likely explanation.

    Finally, you suggest that there is something wrong with disqualifying the credibility of the witnesses. It’s not clear why this is the case. There is such a thing as a difference between reliable and unreliable testimony. There is also good reason to think that biases can render testimony unreliable, a point you are well aware of since you want to cast doubt on the testimony of CEOs and businessmen who are biased by financial incentives. But money is not the only source of bias. If we are to accept the testimony of secondhand smoke subjects, the burden of proof is ours to explain why they can be trusted, in spite of a variety of reasons to think the worst off among them may be biased.

    NS

  2. Dear Noumenal Self,

    thanks for the comment. Allow me to make some clarifications.

    As for the burden question, I’d agree with most of what you’ve said in the first paragraph:

    >Your statement of the burden of proof here is quite dubious. The last I checked, the burden of proof is on anyone making a positive causal statement. In this case, *the positive causal statement would be that secondhand smoke causes health harms.* Surely the burden of proof is on those who allege that this causal relationship exists. Burden of proof is determined by the nature of the claim, not by whether or not it is a claim that CEOs and senators are likely to make.

    I think few could seriously dispute that secondhand smoke causes harm to one’s health. Given that, the burden falls on those who dispute the “settled” question. Burden issues like this–I think at least–are relative to what is considered a settled matter and doesn’t fall universally on the maker of causal claims. If that were the case, all sorts of skeptics could claim absence of causal connection on the most obvious of matters, simply by asserting that the burden always falls on those who assert causal connections. Sometimes the burden falls on those who deny well established causal connections (or things that are considered–perhaps even wrongly–to be well-established causal connections). Considering the mountain of evidence linking smoking and lung cancer, the burden is on the one who disputes that evidence. The case is similar here I think. If, however, Dr.Gori wants to allege that there is no demonstrated causal link between secondhand smoke and specific cancers, that’s fine. He merely needs to point that out–and remind the reader that burden falls on those who make that more narrow assertion.

    Second, I’d also agree with much of what you write here:

    >You criticize this author for ignoring basic knowledge of statistics, yet you never contest his claim that statistical studies are inconclusive about the effects of secondhand smoke. You also never contest his claim that lifetime exposure is never actually measured by these studies. You claim that it is dubious to assert that these measurements are impossible, but in that case you owe us an explanation for why these measurements were not actually taken. If the author is correct that they weren’t, it seems that the impossibility–or at least extreme difficulty–of such measurements could be a likely explanation.

    I don’t contest his claim that those studies are inconclusive. They may be. But he’s alleging that no study could establish the connection between secondhand smoke and lung cancers on account of the difficulty of the measurement. He’s certainly not wrong about the varying conditions of second hand smoke exposure, but those sorts of conditions exist for any variety of phenomena whose properties we deem hazardous. It’s hard, for instance, to measure exposure to poisons in the water–how much, for instance, did a person drink? Whose to say the poison wasn’t dissolved or counteracted by the food they cooked, or by something unknown in the pipes of their house, or by their teeth, or whether it was boiled and so on. In the end, the smoke exposure studies will have to rely on averages of some sort; there are procedures for measuring the accuracy or reliability of those averages. If they are not followed, then fine. But it’s wrong to deny in principle (as Dr.Gori has done here) that such averages are meaningless and wrong.

    And I agree with what you say here–

    >Finally, you suggest that there is something wrong with disqualifying the credibility of the witnesses. It’s not clear why this is the case. There is such a thing as a difference between reliable and unreliable testimony. There is also good reason to think that biases can render testimony unreliable, a point you are well aware of since you want to cast doubt on the testimony of CEOs and businessmen who are biased by financial incentives. But money is not the only source of bias. If we are to accept the testimony of secondhand smoke subjects, the burden of proof is ours to explain why they can be trusted, in spite of a variety of reasons to think the worst off among them may be biased.

    But I’d suggest that good studies have methods for testing observer bias. If they don’t, then Dr.Gori ought to mention that they unquestioningly accepted the testimony of self-interested observers. He suggests this, but doesn’t show it.

    I’d agree finally that it’s good to be skeptical of public science. But there’s a good way to be skeptical and a bad way. Gori’s reasons are bad ones. They are the same kinds of skeptical arguments made by opponents of climate science: skeptical arguments that ask that judgments about trends and averages satisfy impossible conditions of certainty.

    Thanks again for the comment.

  3. Unfortunately, DR Gori is more right than wrong. What is even worse than any dubiouness on either his part or that of the alleged studies is that it is unwise to use bad science to make any kind of policy.

    Personally, I don’t think that DR Gori went far enough in the article.

  4. Maybe someone will be kind enough to explain to me what is meant by “bad science?” Is bad science any result produced which does not turn out to be a law of nature? Is bad science any result that can be contested? If the answer is yes to either of these two latter questions then it would seem that most of quantum mechanics is “bad science.” Do the conditions necessary to call a result bad science vary across different scientific disciplines? Should bad science be defined as the absence of good science, and if so, what qualifies as good science? I’m just looking for a little clarification since I am a bit skeptical that when people talk about good and bad science, they actually know what they are talking about.

    By the way, quantum mechanics is almost exclusively based on probabilities and at a minimum requires a probabilistic understanding of causation. Some have even posited that some quantum events may be entirely uncaused. I can’t help but wonder if some who loosely apply the label of “bad science” would be forced to also call quantum theory the same. However, seeing that you are reading this comment on a device that requires certain aspects of quantum theory to be correct, then that may hint at a problem.
    Note: I am not making a mistake here by applying concepts across both the microscopic and macroscopic realms. I am merely saying that if we call quantum physics science, then it would seem that the same standards about what would be “bad” would apply just as in any other science.

  5. Good question. I remember something Hume said–the wise man proportions his belief to the evidence (or something like that). Even better is something Aristotle said: don’t expect more exactness than is possible. Out of studies as to what causes diseases, for instance, or contributes to them, you’re not going to have mathematical certainty. You might have correlations at best within certain parameters. The smoking industry attack–as “Thank You for Smoking” illustrated–consists in expecting more exactness than is possible for studies such as the secondhand smoking ones. It’s a general strategy. And the generality of Gori’s remarks (none aimed at specific failures of methodology) is a perfect example of that strategy.

  6. Matt — great point! Good thing there isn’t money to be made off doubting the existance of black holes, or conservatives would have started the Center for Black Hole Studies dedicated to attacking “junk science” and the faulty methodology that tried to force “black hole” claptrap on the freedom-loving American public.

  7. Maybe they could have Rush Limbaugh do an offensive mimicking of Stephen Hawking in order to question his authenticity as an astrophysicist.

    JCasey— You gave a solid answer that allows us to preserve the bulk of science and medicine done to day, though the force of my post really wasn’t meant to be directed at you.

  8. Hey MattK–

    I didn’t think your post was directed at me–not that there’s anything wrong with that!–Here’s a link to a New Republic piece about the conservative war on science:

    http://www.tnr.com/doc.mhtml?i=20070129&s=chait012907

    and here’s a highlight:

    >If this sounds like the conservative stance on global warming or evolution, it shouldn’t come as a surprise. Like those two issues, the existence of rising inequality is beyond dispute among academics who study it. This applies even to conservative economists with strong Republican pedigrees. (Harvard economist and former Reagan adviser Martin Feldstein: “There has no doubt been a relatively greater increase in higher incomes in recent years in the United States.” Columbia’s R. Glenn Hubbard, a Bush alum: “We have an issue with emerging inequality in the country.”) And so the ambition of the conservative counterestablishment in these areas is not to overturn the scholarly consensus but simply to make the topic appear so complicated that laypeople and the press don’t know what to believe.

  9. I’m stealing this from the comment section on the original article in the Washington Post:

    >”I wrote a similar opinion in a signed editorial in 1976. It seemed to me that sidestream smoke dissipated in a room should be far less dangerous than a steady stream inhaled directly into the lungs. And I had found a study which purported to demonstrate this. I was wrong. Almost every study done since that time has demonstrated a clear correlation between second-hand smoke and a wide variety of health issues. Scientists have been very cautious about overstating the implications of their conclusions. Every scientific experiment involves some uncertainty in the accuracy of its measurements and scientists have very precise mathematical means for determining how to make up for this: Larger deltas MUST be compensated for by larger sample sizes. Mr Gori is deliberately preying on those who do not understand how statistics are used in science to overcome uncertainty in measurement. While it may seem intuitive that people might have an inexact idea of how much smoke they have been exposed to over a lifetime, the truth is that peoples behavior does segregate them into three broad categories: those who avoid smoke-filled rooms, those who seek them out, and those who dont really care. They report these behaviors pretty accurately and, indeed, can remember with a fairly high degree of precision when they moved from one category to another. If Mr Gori has a critique of the statistics of any particular study or of studies about sidestream smoke in general, he is welcome to make it. That he chooses to misrepresent the basic practices of science in service to his defense of his intuitive beliefs speaks volumes about who is involved in the tyranny of deception.”

    -davemccorkhill

    You can see the rest of the comments (there are many) here: http://www.washingtonpost.com/ac2/wp-dyn/comments/display?contentID=AR2007012901158

  10. Dear jcasey,

    What have you been smoking? The evidence that I am gathering for a position paper very strongly suggests (it could even be said that it screams) that this stuff will kill you. Plus, I can’t find ANYTHING from ANYWHERE that could be considered anywhere in the neighborhood of credible that dissents with that. My advice for you: Pull your head out of your anal orifice.

  11. Dear Stopthemadness,

    Thanks for reading my entry, but I’m having a hard time figuring how you associate me with the false skepticism I am criticizing.

    jc

Comments are closed.