Why the Father of Modern Statistics Didn’t Believe Smoking Caused Cancer

Posted by dnr 4 years, 6 months ago to Science
37 comments | Share | Flag

While Ron Fisher was incorrect about smoking and lung-cancer, he was right about the fact that a controlled study was near impossible and thus there is no correlation/causation proof. The article also makes points about many other topics where there is no proven correlation/causation relationship, one of those topics being climate change.
SOURCE URL: https://priceonomics.com/why-the-father-of-modern-statistics-didnt-believe/


Add Comment

FORMATTING HELP

All Comments Hide marked as read Mark all as read

  • Posted by Herb7734 4 years, 6 months ago
    It cannot be claimed that driving too fast will kill you. Too many variables depending on the skill of the driver, the make and model of the car and other considerations. Yet, for the average rational person, the question becomes, "Why take the chance if it is not necessary?" If you like to speed and you are willing to chance it, go for it, but remember to take into consideration the fact that you might kill others. Smoking is actually similar. Like speeding, not everyone who smokes will develop lung cancer or other respiratory illnesses. There is the factor of harming others with secondary smoke. Again, if you are rational, you won't smoke, but if you love doing it, do it in a way that doesn't harm others. Smoke in open air, speed on race tracks or similar venues. Unfortunately, most people are not fully rational. Rationality in the form of what we used to call common sense is no longer taught or appreciated. So it goes.
    Reply | Mark as read | Best of... | Permalink  
    • Posted by lrshultis 4 years, 6 months ago
      You might update that to no people are fully rational, since, as far as I know, they will go to great extents of mental gymnastics to protect their own beliefs even when confronted with the contrary. I would say that everyone has those little moments when some cherished belief is in danger enough to throw in a little tiny bit of mental dishonesty.
      Decades ago, the second hand smoke study results were somewhat shaky due to two studies showing benefits to secondhand smoke and the evidence not being clear at the standard significance level so the level was reduced to make the studies look significant.
      Reply | Mark as read | Parent | Best of... | Permalink  
      • Posted by Herb7734 4 years, 6 months ago
        In the case of my example the 2nd hand smoke was used to simply complete the comparison. You are right, though, no one is fully rational, not even a computer since they are programed by men. But in using the term rational, it usually means mostly rational. However, that is a clumsy term. I think most people understand that rational when applied to a human being infers the baggage that goes along with it, otherwise we'd need to specify just how rational in every case. 10%, 5%, 15%? You get the point.
        Reply | Mark as read | Parent | Best of... | Permalink  
  • Posted by $ Radio_Randy 4 years, 6 months ago
    Anything that does repeated damage to your body (inside or out) can cause cancer. It's your body's genetic makeup that determines the ease at which you can get cancer.

    Facts:
    My grandfather was a heavy smoker an died, from cancer, in his 80's.
    My grandmother never smoked, yet, died from cancer some years later.
    My mother smoked, for years, but quit. She died, from "lung cancer", some 20 years, later.

    Direct cause, or environmental conditions? Why did my mother die, 20 years after quitting, yet 7 of her children (raised in a smoke filled environment) continue to thrive into their 70's?

    Finally, while the states damn tobacco, they fall over one another to legalize pot. The lack of simple logic astounds me.
    Reply | Mark as read | Best of... | Permalink  
    • Posted by Herb7734 4 years, 6 months ago
      Rarely is anything done by the state logical. That is a mistake that rational people often make because they find it difficult to believe that anyone can progress merrily on not using their faculties.
      Reply | Mark as read | Parent | Best of... | Permalink  
  • Posted by MinorLiberator 4 years, 6 months ago
    Lies, damn lies, and statistics.

    I've never smoked in my life, as I observed as a youth its effects on my parents and other relatives.

    But that was my personal choice.

    In terms of real science, there is no evidence of direct causality, or not nearly enough, to hold tobacco companies legally responsible.

    I also highly doubt the "evidence" of second hand smoke and the legal restrictions mandated.

    To me, that is the same as the much more serious power grab over AGW.
    Reply | Mark as read | Best of... | Permalink  
  • Posted by ProfChuck 4 years, 6 months ago
    Statistical analysis provides a way of characterizing observable behavior. While it may reveal correlations and relationships it does not demonstrate a causal link. Statistics can suggest where to look but it cannot reliably predict what you will find when you do so. Statistical analysis is also subject to hidden assumptions. For example, there is a strong correlation between shoe size and mathematical ability. Now this may seem absurd at first but when you realize that a representative sample must include infants it makes sense.
    Reply | Mark as read | Best of... | Permalink  
  • Posted by Leonid 4 years, 6 months ago
    A scientific proof of causation requires observation of cause effect connection with certainty which can be independently reproduced. When we use statistics instead we have such a bizzare causes of cancer like cellphone microwave high voltage lines or what have you. The last example of this anti-scientific approach is a claim that vaccines cause autism. Statistically people with blue eyes have much more heart attacks than with brown. Can we really correlate the color of eyes with heart disease?
    Reply | Mark as read | Best of... | Permalink  
  • Posted by Leonid 4 years, 6 months ago
    A scientific proof of causation requires observation of cause effect connection with certainty which can be independently reproduced. When we use statistics instead we have such a bizzare causes of cancer like cellphone microwave high voltage lines or what have you. The last example of this anti-scientific approach is a claim that vaccines cause autism. Statistically people with blue eyes have much more heart attacks than with brown. Can we really correlate the color of eyes with heart disease?
    Reply | Mark as read | Best of... | Permalink  
  • Posted by 4 years, 6 months ago
    For me the most important thing that this article identifies is that many things do not have correlation/causation proof. Specifically, I thought that with climate change we have the same issue as with smoking. We cannot do a controlled studies of one where we are not emitting any CO2 and one where we do. So, what do we do? We build models. For example, one climate model uses as an initial condition the assumption of CO2 doubles in say 50 years and then they run the model. The model assumes that there is causation between CO2 levels and temperature. The problem is that this is an assumption. There is no proof to substantiate the assumption. Now maybe, like with smoking, there will be an overwhelming preponderance of evidence that leads us all to believe that such a linkage exists, but we don't have it. It could be that CO2 decreases world-wide temperature. Right now we don't know.
    Reply | Mark as read | Best of... | Permalink  
    • Posted by lrshultis 4 years, 6 months ago
      CO2 in the upper troposphere is likely the way the atmosphere releases radiation to space because the 99% of gases of the atmosphere N2, O2, and Ar do not radiate well but hold energy and thus have temperature which has to be radiated to space. CO2 can be heated by them and then radiate the energy away and then with convection working, the lapse rate of about 6 deg C decrease per kilometer comes about. At least that is the way I understand it.
      Reply | Mark as read | Parent | Best of... | Permalink  
  • Posted by DrZarkov99 4 years, 6 months ago
    Genetics do have a lot to do with susceptibility to cancer. My family history going back a couple of centuries has zero incidence of any kind of cancer or heart disease. Both my parents smoked until the last few years of their life, and died from unrelated disease in their 80s. However, my kind of stock appears to be rare.
    Reply | Mark as read | Best of... | Permalink  
  • Posted by LibertyBelle 4 years, 6 months ago
    I am not a statistician. I wild admit that the burden of proof properly rests on the side that makes the positive assertion (for instance "This causes that.")
    --However, whatever evidence exists in favor of some foods/drinks versus others, the fact is that
    smoke is neither food nor drink; it does not natu-
    rally belong in the body, and its initial introduction is likely to cause coughing, or other
    physical discomfort. It is really unwise to get
    into such a habit. (Not to mention the stench,
    extremely offensive to a non-practicioner). I can
    see why an addict would have a motivation to
    convince people (or primarily, himself) that it is
    not harmful, but I have to disagree (with all due
    respect to anyone, whatever that person's a-
    chievements, who tried to glamorize it).
    Reply | Mark as read | Best of... | Permalink  
  • Posted by $ Abaco 4 years, 6 months ago
    A good friend of mine used to deny the risks of smoking. Then, he admitted that he quit when they were wheeling him into the cath lab for angioplasty while the nurse explained how smoking makes your blood sticky. Luckily, my friend is still alive...and is a real character.
    Reply | Mark as read | Best of... | Permalink  
  • Posted by $ blarman 4 years, 6 months ago
    That's all nice and good, but all one has to do is look at the lung tissue from someone who has smoked for years and compare it to someone who hasn't to see pretty clearly that smoking isn't good for you. One can rail on and on about causation (or lack thereof), but the physical evidence is there in the lungs. I think what is amazing is that people who quit smoking heal - even if they smoked for years. It can take months to completely rebuild all the alveoli destroyed by smoking, but it can happen. But what strikes me is that if the body is repairing damage, how someone would discount this reality in an attempt to downplay the very real dangers of smoking (which do include cancer) is to me avoidance of reality.
    Reply | Mark as read | Best of... | Permalink  
  • Posted by Leonid 4 years, 6 months ago
    Re-Tobacco companies and second hand smoking. It's a political, not scientific matter. It was very important step in the modern trait of vilification of the business to present capitalists as profiting from illness and death. It was also important statist further curbing of property rights by regulation of smoking in private establishments.
    Reply | Mark as read | Best of... | Permalink  
  • Posted by jdg 4 years, 6 months ago
    Anyone who tries to prove that climate change is or is not happening is attacking a straw man. Everyone knows the climate and temperature change over geological ages. The important questions to ask about climate change are (1) is it likely to have any ill effects on humans, or on anything that humans depend on? (2) If yes, what is the most effective, cheapest, and fastest way to stop it or alleviate the ill effects?

    So far, Gregory Benford's boatload of iron filings appears a much better answer to (2) than the Al Gore program of cutting back on energy use. But even that may very well be unnecessary, because no one on either side of the debate has even tried to answer (1) yet.
    Reply | Mark as read | Best of... | Permalink  
    • Posted by Herb7734 4 years, 6 months ago
      Anything put forth by Algore should be completely discounted. His advocacy of "climate change" is motivated by self-aggrandizement, and a way to make large sums of money. I have no complaint with making large sums of money, but not by promoting a fraud.
      Reply | Mark as read | Parent | Best of... | Permalink  
  • Posted by Leonid 4 years, 6 months ago
    Statistics is never proof for anything. It's applicable only to the large group of people give you only probability that X causes Y. In other words we are talking about potential, not actual reality. Smoking causes certain kind of lung cancer in 1 out of 800 people. So probability is 0,125 %. But that doesn't apply to the real .smoker. For him is always 100 or 0%.So how one can establish cause effect connection based on probability?
    Reply | Mark as read | Best of... | Permalink  
    • Posted by 4 years, 6 months ago
      I think that causation is different that just a probability. When you can link correlation to causation then you have a proof. Not a formal as a mathematical proof, but can be close.
      Reply | Mark as read | Parent | Best of... | Permalink  
  • Posted by Eyecu2 4 years, 6 months ago
    I agree that eliminating all possible factors is impossible and therefore causality is not provable. However I am sure that smoking is not a healthy thing to do.
    Reply | Mark as read | Best of... | Permalink  
  • Posted by $ Olduglycarl 4 years, 6 months ago
    It's the poisons they put in cigs that causes problems and Yes, some are more vulnerable to anything but clean air in their lungs. But there are those that think that a virus and pathogens are the real cause of lung cancer. (true, smoking is not a good thing but not a direct cause).
    The causes of cancer now are being re-thought finally...accounting for high levels of heavy toxic metals, pesticides etc, candida yeast overrun and a poor immune system, Which anyone on prescription drugs and have been vaccinated to death would fit into that category.

    Cancer is such a money maker for the lamestream allopathic sickness system...it's amazing anyone has been able to see thru the bull crap.
    Reply | Mark as read | Best of... | Permalink  
    • Posted by $ sjatkins 4 years, 6 months ago
      The correlation between smoking and lung cancer is so high that few would argue there is no causation.
      Reply | Mark as read | Parent | Best of... | Permalink  
      • Posted by 4 years, 6 months ago
        There cannot be causation unless you have actual studies that confirm it. No such study was done (or could be done) re smoking and cancer. The subjective proof is clear, but that is not a formal statistical proof.
        Reply | Mark as read | Parent | Best of... | Permalink  

FORMATTING HELP

  • Comment hidden. Undo