4 ms·

I think you’re confusing the map for the territory. The study pooled all workers together, but that doesn’t invalidate the hypothesis that certain types of work

I think you’re confusing the map for the territory. The study pooled all workers together, but that doesn’t invalidate the hypothesis that certain types of workers were more affected. Unless the study explicitly broke out the purported subdivision, we can’t say for sure. For example imagine 50% of workers are office workers. And say heart attack risk is twice as high for overworked office workers but 10% less for non office workers. The study would find a broad-based 40% increase in risk for all workers, despite OPs original hypothesis being exactly correct.

WhompingWindows 5y ago
The parent comment shoots from the hip, though. He has one anecdotal case of a more active career living to 92, then he postulates his inactive career will lead him to an extremely early death (60 is young in modern economies). That's just not the depth of discussion HN is designed for, does this ignite curiosity or push the discussion further?
- throwaway8581 5y ago
  Yes, it does. This is exactly how you pick holes at research. What factors didn't it look at? That helps us understand the limits of what we can deduce from the data. Can we think of an intuitive story about why those factors might have a causal relationship to the thing being measured? If so, we have especially good reason to be skeptical of any conclusions from the study that don't address that possible explanation. Somewhere in the last couple of decades, the SCIENCE WORKS, BITCHES people seem to have forgotten how real science is actually done. Intuition, anecdote, common sense, hunches are very important parts of the process.
  - WhompingWindows 5y ago
    Science works that way in the minds of scientists, whose vast experience gives them dozens of actual data studies they've read to speculate from. Researchers have read about thousands of cases, they're the ones with the refined intuition, not forum readers. Specifically: Is it productive for HN users to have anecdotal evidence rise to the top, or actual peer-reviewed evidence from studies with N=100 or N=10,000? I'd argue if layman HN users have time to read only one comment, we should upvote the N>1 comments, not N=1. From the guidelines: "Comments should get more thoughtful and substantive, not less, as a topic gets more divisive." We went from an international article (N=745k) to anecdata, which is a downgrade of substance in my view.
freddie_mercury 5y ago
I'm not sure what's the point of coming up with farfetched theoreticals disconnected from the actual study when it is transparently obvious the OP was just reacting to a headline with a preconceived notion. They found the biggest effects in Southeast Asia. I live in Southeast Asia. If you think office workers make up a significant part of workers in Southeast Asia....That's just not a "let's assume" I'm going to join you with. If you think office workers are the cause then do the hard work of actually looking at the data and use that to make your case.

WhompingWindows 5y ago

The parent comment shoots from the hip, though. He has one anecdotal case of a more active career living to 92, then he postulates his inactive career will lead him to an extremely early death (60 is young in modern economies). That's just not the depth of discussion HN is designed for, does this ignite curiosity or push the discussion further?

throwaway8581 5y ago
Yes, it does. This is exactly how you pick holes at research. What factors didn't it look at? That helps us understand the limits of what we can deduce from the data. Can we think of an intuitive story about why those factors might have a causal relationship to the thing being measured? If so, we have especially good reason to be skeptical of any conclusions from the study that don't address that possible explanation. Somewhere in the last couple of decades, the SCIENCE WORKS, BITCHES people seem to have forgotten how real science is actually done. Intuition, anecdote, common sense, hunches are very important parts of the process.
- WhompingWindows 5y ago
  Science works that way in the minds of scientists, whose vast experience gives them dozens of actual data studies they've read to speculate from. Researchers have read about thousands of cases, they're the ones with the refined intuition, not forum readers. Specifically: Is it productive for HN users to have anecdotal evidence rise to the top, or actual peer-reviewed evidence from studies with N=100 or N=10,000? I'd argue if layman HN users have time to read only one comment, we should upvote the N>1 comments, not N=1. From the guidelines: "Comments should get more thoughtful and substantive, not less, as a topic gets more divisive." We went from an international article (N=745k) to anecdata, which is a downgrade of substance in my view.

freddie_mercury 5y ago

I'm not sure what's the point of coming up with farfetched theoreticals disconnected from the actual study when it is transparently obvious the OP was just reacting to a headline with a preconceived notion. They found the biggest effects in Southeast Asia. I live in Southeast Asia. If you think office workers make up a significant part of workers in Southeast Asia....That's just not a "let's assume" I'm going to join you with. If you think office workers are the cause then do the hard work of actually looking at the data and use that to make your case.