5 ms·

Correlation (people who ate peppers died less), in no way implies their stated causation ( eating peppers makes you live longer ). Consider how many old people

Correlation (people who ate peppers died less), in no way implies their stated causation ( eating peppers makes you live longer ). Consider how many old people vs young people like spicy food and I think you get my point. I really just wanted some damned confirmation bias, not shitty reporting.

kekebo 10y ago
There was a larger study done in china a few years earlier with more differentiated results and the correlations are indeed interesting: http://www.bmj.com/content/351/bmj.h3942 http://www.bmj.com/content/351/bmj.h3942 . But yeah, not a good article.
gilrain 10y ago
Why do you think these researchers were so incompetent that they didn't control for even blindingly-obvious factors like age? Are you familiar with their other work in the field, or something? "Correlation does not equal causation" is parroted more often by pseudo-intellectuals than scientists, now. It seems that many interpret it as "correlation is not causation and is therefore garbage which can tell us nothing".
- placeybordeaux 10y ago
  Correlation can tell us where to search, but the article concludes with the author being smug about sprinkling chili onto his omelette. There are still plenty of confounding factors, hardest of which to control for post-hoc would be self selection bias. This study just takes the data from a prior study and analyzed to see if one of the specific variables would predict another. There's no way to tell if they just did this for all of the variables, if they did this without having a prior hypothesis then the results should only be used for constructing another study, not as interpretable results that we can take action on.
ctdonath 10y ago
Correlation implies causation, which is why scientists proceed from there in search of proof. You're trying to contend it doesn't prove, and thereby imply the whole of the research involved is wrong. Incoherent accusations, poor punctuation, and profanity correlate with incompetence. They don't prove you are, but you certainly might be.
- dragonwriter 10y ago
  > Correlation implies causation Strictly logically speaking, it does not. Probabilistically, it does; or, IOW, correlation correlates with causation.
  - ctdonath 10y ago
    correlation correlates with causation I like that.
  - fragsworth 10y ago
    Actually, probabilistically, it does not. There's effectively infinite correlations you can come up with that have no cause and effect
    - dragonwriter 10y ago
      When I say there is a probabilistic implication, I mean, basically: P (X causes Y|X correlates with Y) > P (X causes Y) > P (X causes Y|X doesn't correlate with Y)
      - fragsworth 10y ago
        Yeah, I don't think that conjecture holds (or even makes sense?), if you're talking about the set of all possible correlations. If you take the limits, you end up getting 0 > 0 > 0.
- placeybordeaux 10y ago
  > Correlation implies causation It doesn't. > which is why scientists proceed from there in search of proof. Scientists don't prove anything, in fact all they do is disprove or fail to do so.
misja111 10y ago
The article has a link to the original research paper. If you read it, you find: > Total mortality for participants who consumed hot red chili peppers was 21.6% compared to 33.6% for those who did not (absolute risk reduction of 12%; relative risk of 0.64). Adjusted for demographic, lifestyle, and clinical characteristics, the hazard ratio was 0.87.

kekebo 10y ago

There was a larger study done in china a few years earlier with more differentiated results and the correlations are indeed interesting: http://www.bmj.com/content/351/bmj.h3942 http://www.bmj.com/content/351/bmj.h3942 . But yeah, not a good article.

gilrain 10y ago

Why do you think these researchers were so incompetent that they didn't control for even blindingly-obvious factors like age? Are you familiar with their other work in the field, or something? "Correlation does not equal causation" is parroted more often by pseudo-intellectuals than scientists, now. It seems that many interpret it as "correlation is not causation and is therefore garbage which can tell us nothing".

placeybordeaux 10y ago
Correlation can tell us where to search, but the article concludes with the author being smug about sprinkling chili onto his omelette. There are still plenty of confounding factors, hardest of which to control for post-hoc would be self selection bias. This study just takes the data from a prior study and analyzed to see if one of the specific variables would predict another. There's no way to tell if they just did this for all of the variables, if they did this without having a prior hypothesis then the results should only be used for constructing another study, not as interpretable results that we can take action on.

ctdonath 10y ago

Correlation implies causation, which is why scientists proceed from there in search of proof. You're trying to contend it doesn't prove, and thereby imply the whole of the research involved is wrong. Incoherent accusations, poor punctuation, and profanity correlate with incompetence. They don't prove you are, but you certainly might be.

dragonwriter 10y ago
> Correlation implies causation Strictly logically speaking, it does not. Probabilistically, it does; or, IOW, correlation correlates with causation.
- ctdonath 10y ago
  correlation correlates with causation I like that.
- fragsworth 10y ago
  Actually, probabilistically, it does not. There's effectively infinite correlations you can come up with that have no cause and effect
  - dragonwriter 10y ago
    When I say there is a probabilistic implication, I mean, basically: P (X causes Y|X correlates with Y) > P (X causes Y) > P (X causes Y|X doesn't correlate with Y)
    - fragsworth 10y ago
      Yeah, I don't think that conjecture holds (or even makes sense?), if you're talking about the set of all possible correlations. If you take the limits, you end up getting 0 > 0 > 0.
placeybordeaux 10y ago
> Correlation implies causation It doesn't. > which is why scientists proceed from there in search of proof. Scientists don't prove anything, in fact all they do is disprove or fail to do so.

misja111 10y ago

The article has a link to the original research paper. If you read it, you find: > Total mortality for participants who consumed hot red chili peppers was 21.6% compared to 33.6% for those who did not (absolute risk reduction of 12%; relative risk of 0.64). Adjusted for demographic, lifestyle, and clinical characteristics, the hazard ratio was 0.87.