5 ms·

"Do Cigarette Warning Labels Help CAUSE Cancer?" No... Long answer: Hell no and stop doing this. To the next person with a pseudoscientific gut-feeling-bear-wi

by eksith 14y ago

"Do Cigarette Warning Labels Help CAUSE Cancer?" No... Long answer: Hell no and stop doing this. To the next person with a pseudoscientific gut-feeling-bear-with-me-just-a-bit-while-I-pull-a-half-baked-idea-out-of-a-number-set-that-does-nothing-to-corroborate, please don't. Corrolation doesn't equate to causation, no matter what amount of reality bending assertions you apply. Warning those at high risk for cancer know they're at high risk and subsequently confirm the hypothesis doesn't mean the warning was the cause.

zzzeek 14y ago
There seems to be a trend of Very Smart People with no experience whatsoever in a given field, reading some article somewhere, pulling out R and some graphing software, and proclaiming "I've discovered something that nobody who actually works in the target field which I knew absolutely nothing about six hours ago ever found!". I wish those stories would stop getting posted to HN.
- stfu 14y ago
  NO! I absolutely resent that idea that "science" should only be something for the experts, and everybody without a PhD should shut up about it. As long as it is just random ranting without any data, it is more than welcoming that "regular" people try to understand expert opinion and in fact challenge it. If "facts" are so clear and indisputable, than those amateur research should easily be shown as wrong. If not, amateur researcher make a highly valuable contribution. They highlight the fact, that social science, especially on social phenomenons, is (often) based on consensus. In this case here, the author makes an excellent contribution just by highlighting how easy correlation and causation can be (mis)interpreted. For example the other day I read and article published in a well regarded management journal, by authors who claim having found out that the willingness of users clicking on ads is based on the hours of sunshine per day in that region. The correlation equals causation phenomenon is quite common among social scientists who try to show how scientific their contributions are.
  - eksith 14y ago
    You're right, science shouldn't just be reserved for the PhDs. But the PhD is likely to have insight and experience that will make a lot of assertions false, avenues of research redundant and generally be able to tell the different between science and not science. There's no restriction on laymen studying, researching and trying to understand new frontiers of science, however just as the PhDs don't have a monopoly on science (and as you point out, they're far from perfect), neither do lay people have a monopoly on reinterpreting reality through the misinterpretation of facts. tl;dr: You're entitled to examination of facts and to be examined for your own research. You're, however, not entitled to your own reality. Which is what happened in the article.
  - zzzeek 14y ago
    > I absolutely resent that idea that "science" should only be something for the experts, and everybody without a PhD should shut up about it. making some off the cuff, armchair observations within a particular field of study (and blogging and serially promoting them to hacker news) without even making casual efforts to familiarize oneself with at least portions the vast amounts of knowledge publicly available on that field is, generously, pretty bad "science". And less generously, it's just hubris. This is not an "amateur researcher", this is someone who apparently based on posting history has a really high blogging-to-value ratio. Of course, we all have the right to be clueless. I only requested that perhaps Hacker News doesn't want to encourage a culture of armchair cluelessness and particularly someone who appears to be a serial armchair blogger/HN promoter (see https://news.ycombinator.com/submitted?id=techdog https://news.ycombinator.com/submitted?id=techdog).
- dfc 14y ago
  That graph did not look nice enough to come from R.
- tokenadult 14y ago
  There seems to be a trend of Very Smart People with no experience whatsoever in a given field, reading some article somewhere, pulling out R and some graphing software, and proclaiming "I've discovered something that nobody who actually works in the target field which I knew absolutely nothing about six hours ago ever found!". I wish those stories would stop getting posted to HN. I've seen one of these (on a different topic) posted to HN repeatedly in comments, by a variety of participants here. Really, people need to learn how to evaluate sources better here on Hacker News. That's part of being an educated person. http://news.ycombinator.com/item?id=5105664 http://news.ycombinator.com/item?id=5105664
- robbiep 14y ago
  I feel it is the hacker mentality. People here are very intelligent but (some often) make the mistake of feeling as though their experience and success in one domain is applicable to multiple, that they are, in fact, renaissance men and women when in fact they are largely unskilled in the Domains they are attempting to analyse in depth. As someone heavily involved in medicine j see this all the time when the subject matter is something medical or biological- people really so feel that they can talk about issues of medicine and science with just the same authority as they can when otherwise discussing programming/whatever their day job is, when in fact they dont know just how much they don't know
- randomdata 14y ago
  I found it to be thought provoking, even if it wasn't able to form any definite conclusions. Correlation is still interesting for many reasons. Your post suggests that it has been researched before (and presumably proved to show no link), but searching Google did not yield any results for me. I, and I'm sure most everyone who has read this article, would love to look over the findings if you are able to share what you know.
  - zzzeek 14y ago
    The blog post in question states in the fourth paragraph he was inspired by the recent article about "placebos and nocebos", which is "the article" I'm referring to. That is, we're reading the product of someone's short-term random curiosity. Random curiosity is a great thing but IMHO the path from random curiosity to published findings needs to have some serious experience in the middle for it to be worth our time.
Zigurd 14y ago
If he had stopped at "Hmmm, there may be some cause or causes of lung cancer we're not paying enough attention to." that would have been fine.
- eksith 14y ago
  Agreed. There's a big difference between "I don't know" and "I don't know, therefore...". People jump to therefore based on assumptions and that's not science. The next step after "I don't know" should be "let's investigate objectively".
rmc 14y ago
Corrolation doesn't equate to causation Yes this is correct. And to go off on a general tangent, correlation tells you where to start looking. After all, if A & B are correlated, then one of these statements must be true: "A causes B", "B causes A" or "there is something else, C, which causes A and B". Just because you can find a correlation, doesn't mean that finding is irrelevant. It is very interesting and tells you where to start looking.
- ceejayoz 14y ago
  > After all, if A & B are correlated, then one of these statements must be true: "A causes B", "B causes A" or "there is something else, C, which causes A and B". You missed "it's purely coincidence", which is an entirely valid option.
  - rmc 14y ago
    You should do enough experiments to be able to state with statistical confidence that it's not a coincidence.
    - greenyoda 14y ago
      The correlation could also be due to a flaw in your experimental design (e.g., something in your lab is contaminating your samples, or the researcher has an unconscious bias that skews measurements in a particular direction), in which case repeating the same experiments will not solve the problem.
      - rmc 14y ago
        Sure, there's loads of ways an experiment can be flawed. But also, it might not be flawed. If you're got a good experiment that shows correlation, then it has to be one of the 3 results above.
        ceejayoz 14y ago
        No, it is statistically most likely to be one of the 3 results above. "Has to be" doesn't have much of a place in a statistics discussion. There is a non-zero probability of the correlation being due to chance.

by eksith 14y ago

zzzeek 14y ago

There seems to be a trend of Very Smart People with no experience whatsoever in a given field, reading some article somewhere, pulling out R and some graphing software, and proclaiming "I've discovered something that nobody who actually works in the target field which I knew absolutely nothing about six hours ago ever found!". I wish those stories would stop getting posted to HN.

stfu 14y ago
NO! I absolutely resent that idea that "science" should only be something for the experts, and everybody without a PhD should shut up about it. As long as it is just random ranting without any data, it is more than welcoming that "regular" people try to understand expert opinion and in fact challenge it. If "facts" are so clear and indisputable, than those amateur research should easily be shown as wrong. If not, amateur researcher make a highly valuable contribution. They highlight the fact, that social science, especially on social phenomenons, is (often) based on consensus. In this case here, the author makes an excellent contribution just by highlighting how easy correlation and causation can be (mis)interpreted. For example the other day I read and article published in a well regarded management journal, by authors who claim having found out that the willingness of users clicking on ads is based on the hours of sunshine per day in that region. The correlation equals causation phenomenon is quite common among social scientists who try to show how scientific their contributions are.
- eksith 14y ago
  You're right, science shouldn't just be reserved for the PhDs. But the PhD is likely to have insight and experience that will make a lot of assertions false, avenues of research redundant and generally be able to tell the different between science and not science. There's no restriction on laymen studying, researching and trying to understand new frontiers of science, however just as the PhDs don't have a monopoly on science (and as you point out, they're far from perfect), neither do lay people have a monopoly on reinterpreting reality through the misinterpretation of facts. tl;dr: You're entitled to examination of facts and to be examined for your own research. You're, however, not entitled to your own reality. Which is what happened in the article.
- zzzeek 14y ago
  > I absolutely resent that idea that "science" should only be something for the experts, and everybody without a PhD should shut up about it. making some off the cuff, armchair observations within a particular field of study (and blogging and serially promoting them to hacker news) without even making casual efforts to familiarize oneself with at least portions the vast amounts of knowledge publicly available on that field is, generously, pretty bad "science". And less generously, it's just hubris. This is not an "amateur researcher", this is someone who apparently based on posting history has a really high blogging-to-value ratio. Of course, we all have the right to be clueless. I only requested that perhaps Hacker News doesn't want to encourage a culture of armchair cluelessness and particularly someone who appears to be a serial armchair blogger/HN promoter (see https://news.ycombinator.com/submitted?id=techdog https://news.ycombinator.com/submitted?id=techdog).
dfc 14y ago
That graph did not look nice enough to come from R.
tokenadult 14y ago
There seems to be a trend of Very Smart People with no experience whatsoever in a given field, reading some article somewhere, pulling out R and some graphing software, and proclaiming "I've discovered something that nobody who actually works in the target field which I knew absolutely nothing about six hours ago ever found!". I wish those stories would stop getting posted to HN. I've seen one of these (on a different topic) posted to HN repeatedly in comments, by a variety of participants here. Really, people need to learn how to evaluate sources better here on Hacker News. That's part of being an educated person. http://news.ycombinator.com/item?id=5105664 http://news.ycombinator.com/item?id=5105664
robbiep 14y ago
I feel it is the hacker mentality. People here are very intelligent but (some often) make the mistake of feeling as though their experience and success in one domain is applicable to multiple, that they are, in fact, renaissance men and women when in fact they are largely unskilled in the Domains they are attempting to analyse in depth. As someone heavily involved in medicine j see this all the time when the subject matter is something medical or biological- people really so feel that they can talk about issues of medicine and science with just the same authority as they can when otherwise discussing programming/whatever their day job is, when in fact they dont know just how much they don't know
randomdata 14y ago
I found it to be thought provoking, even if it wasn't able to form any definite conclusions. Correlation is still interesting for many reasons. Your post suggests that it has been researched before (and presumably proved to show no link), but searching Google did not yield any results for me. I, and I'm sure most everyone who has read this article, would love to look over the findings if you are able to share what you know.
- zzzeek 14y ago
  The blog post in question states in the fourth paragraph he was inspired by the recent article about "placebos and nocebos", which is "the article" I'm referring to. That is, we're reading the product of someone's short-term random curiosity. Random curiosity is a great thing but IMHO the path from random curiosity to published findings needs to have some serious experience in the middle for it to be worth our time.

Zigurd 14y ago

If he had stopped at "Hmmm, there may be some cause or causes of lung cancer we're not paying enough attention to." that would have been fine.

eksith 14y ago
Agreed. There's a big difference between "I don't know" and "I don't know, therefore...". People jump to therefore based on assumptions and that's not science. The next step after "I don't know" should be "let's investigate objectively".

rmc 14y ago

Corrolation doesn't equate to causation Yes this is correct. And to go off on a general tangent, correlation tells you where to start looking. After all, if A & B are correlated, then one of these statements must be true: "A causes B", "B causes A" or "there is something else, C, which causes A and B". Just because you can find a correlation, doesn't mean that finding is irrelevant. It is very interesting and tells you where to start looking.

ceejayoz 14y ago
> After all, if A & B are correlated, then one of these statements must be true: "A causes B", "B causes A" or "there is something else, C, which causes A and B". You missed "it's purely coincidence", which is an entirely valid option.
- rmc 14y ago
  You should do enough experiments to be able to state with statistical confidence that it's not a coincidence.
  - greenyoda 14y ago
    The correlation could also be due to a flaw in your experimental design (e.g., something in your lab is contaminating your samples, or the researcher has an unconscious bias that skews measurements in a particular direction), in which case repeating the same experiments will not solve the problem.
    - rmc 14y ago
      Sure, there's loads of ways an experiment can be flawed. But also, it might not be flawed. If you're got a good experiment that shows correlation, then it has to be one of the 3 results above.
      - ceejayoz 14y ago
        No, it is statistically most likely to be one of the 3 results above. "Has to be" doesn't have much of a place in a statistics discussion. There is a non-zero probability of the correlation being due to chance.