7 ms·

Weird that people seem to assert that causality is not established? There are hundreds to thousands of studies that show that pretty much all markers for morta

Weird that people seem to assert that causality is not established? There are hundreds to thousands of studies that show that pretty much all markers for mortality and morbidity go down with excercise, i.e. put someone regulalrly on a treadmill for a couple of month and pretty much all health markers improve. People implying that the study shows only correlation really don't seem to understand how we establish causality in science.

prerok 2y ago
Well, I don't doubt that cardio fitness decreases the chance for morbidity and mortality, but in this study they only established correlation. I think that in science we have to be extra careful just what we are saying. And saying that this study proves causality would be wrong. I am sure there are other studies that could (do?) prove it but this statistical analysis can only prove correlation.
- bigfudge 2y ago
  We only have correlation for lots of important results. Smoking cigarettes is one example. Experiments are almost always better, and a lot of correlational research is terrible. But knowing something for sure doesn’t always require an experiment.
  - caeril 2y ago
    That's a bit different. There's no reasonable path to reverse-causality with smoking. "People prone to developing lung cancer are more likely to take up smoking" doesn't really make sense. On the other hand, "Fat and unhealthy people are more prone not to exercise" is a reasonable reverse-causality position, for all kind of reasons. Your knees hurt, you're self-conscious, it's extremely uncomfortable, your fat jiggles, etc. I'm not saying direct causality doesn't exist here, I'm just saying that reverse-causality is reasonable, whereas with smoking it is not. Assuming causation from this particular correlation is harder than it is with smoking.
    - lukas099 2y ago
      "People who smoke are more likely to do other things that the medical establishment condemns" is a plausible non-causal explanation for the correlation between smoking and lung cancer. I think that we just have to look at the best explanation for the current data to determine causality, but it’s always a guess.
      - dllthomas 2y ago
        That is a plausible explanation for a present correlation, but I think less so for when the correlation was first noted.
- tea-coffee 2y ago
  From the Methods section of the paper (under "study design"): "We also included meta-analyses that pooled data from primary prospective/retrospective cohort or case-control studies. These studies were the focus because of their ability to assess causality for observational research". While "causality" is a strong word in the sentence above, the data is much stronger than simple correlation. Of course, outright causality has not been established, but the evidence to determine a predictive association is strong.
Fricken 2y ago
Exercise is a four letter word for many on HN. Content posted here touting the various benefits of exercise is always met with mental gymnastics from commenters who would rather do anything but go out and break a sweat.
- portaouflop 2y ago
  I used to be like that- for some reason I started exercising this year and it completely changed my life. I feel better than ever and it’s not like I’m training a lot, a couple of times per week. If you’re on the fence and reading this definitely give it another shot
  - Eddy_Viscosity2 2y ago
    Second this, it doesn't just help your body. Your mental health will also see benefits.
- agumonkey 2y ago
  Tiny side note in case someone is interested: my health crashed in 2014, I used to be a runner, now I can't jog nor reach sweat levels of efforts. But you don't need to reach that level to restore/improve your health tremendously. - a tiny step to do small one leg squat - a kettle bell or any weight for deadlift doing 10 repetition at low speed, low intensity[0] of the two abose everytime you get bored, anxious, or lost on a youtube rabbithole will feel like nothing, yet, over the week you'll start to feel muscle grow slowly, less joint pain, better posture, better ability to move, everything will feel easier thus having a better mood too. once in the morning, once in the evening.. or maybe more as one sees fit, until you feel the drive to do more (usually a month of slow and pleasurable exercise will naturally lead to a desire to try more) [0] real slow, like taichi slow.. no muscle burn, no fast breathing..
- Workaccount2 2y ago
  For HN's that want to get active, I cannot suggest getting a (higher tier) Garmin watch enough. I know you have the money so just drop the $800. It's stats on stats on stats and essentially creates a "character health" panel for your own body. Charts, metrics, real time tracking and exportable data. There are achievement badges and social connections if that's your thing too (good for you and your SO). For me, being able to see real-time metrics and real time improvement really amped up my motivation to go out an exercise. Rewards were no longer this mysterious ethereal thing that will maybe show up sometime in the future if I keep grinding. Day by day, I could see improvement and sure enough could feel it too.
huijzer 2y ago
No sorry but this is not how causality works. When event B happens a thousand times after event A, you cannot conclude that A causes B. For example, there could be a confounder, that is, a third event that triggers A and then B (so not A triggers B, but this third event triggers B). See The Book of Why by Pearl for more information on causality.
- gjjydfhgd 2y ago
  You cannot conclude, but at the same time you should increase your probability of causality. Keeping it unchanged is another kind of logical error.
- bigfudge 2y ago
  But Pearls book explicitly argues for using correlational evidence to infer causation. The trick is to actually have a causal model and design observational studies that rule out as many alternative explanations as possible.
  - cycomanic 2y ago
    Thank you! That seems to be the issue with many commenter here, they believe you can have a single experiment that somehow proofs causality. The same scepticism would not be brought against many "hard science" experiments, even though they do the exact same thing, falsify alternative explanations until they have high certainty that they have causality.
- knallfrosch 2y ago
  Can you give an example where causality is established then? It sounds like I can hit you a 1,000 times, you feel pain a 1,000 times and you still don't believe there's a causality.
  - gwd 2y ago
    Consider this quote: "Every time we send 5 fire trucks to a fire, the damage is 10x than when we only send one fire truck. We've observed this 1000 times. And still you don't believe that the fire trucks are causing the damage." In this case it should be absolutely clear that A (lots of trucks) aren't causing B (lots of damage), but rather a third aspect, C (size of fire) is causing both A and B. Insisting that A causes B will result in completely counterproductive interventions, like "send only one truck to all fires". The same thing could be true for cardiovascular fitness. If people are sick, they're much less likely to running or hiking up a mountain. So rather than poor cardio fitness (A) causing high mortality (B), it could be that a third thing, sickness (C) is causing both A and B. If that is the case, then shaming people who are sick into trying to exercise, instead of making them healthy enough so that they feel like running, is likely to make things worse rather than better. How do you tell the difference? Well the "gold standard" is randomized controlled trials. Pick 3,000 random people. Tell 1000 of them to exercise more, and 1000 of them to exercise less, and 1000 leave alone, and compare. If the "exercise more" group is healthier at the end of 10 years, that's decent evidence that "exercise more" is a useful intervention. Failing that, you can think of other possible confounding factors and control for them. Don't just ask how much they exercise; ask how old they are, and how well they are, how stressed they are, and loads of other factors which might both cause both A and B, and use statistical methods to detect whether one of those factors is actually a better explanation than "A -> B".
bjornsing 2y ago
> People implying that the study shows only correlation really don't seem to understand how we establish causality in science. Or perhaps you don’t fully understand the challenges of establishing causality? Just because an intervention causes an improvement in some bio markers that are associated with lower mortality (unfortunately) does not mean that the intervention will cause lower mortality. The classic example is vitamin D supplements: Higher vitamin D levels are associated with lower mortality in many medical conditions. Vitamin D supplements increase vitamin D levels. But vitamin D supplements seldom lower mortality. Why? Probably because vitamin D is produced in the skin when we are in the sun. The more healthy subpopulation of any study will typically spend more time outside, so they will have higher vitamin D levels. But it’s (relative) health that causes higher vitamin D, not the other way around. The only way to reliably establish causality is really an end-to-end randomized controlled trial (RCT). Stitching together two RCTs is not sufficient. (Not saying that exercise does not lower mortality BTW, just that it’s complicated, and a study such as this is probably picking up two signals: one causal and one purely correlational.)
- cycomanic 2y ago
  > > People implying that the study shows only correlation really don't seem to understand how we establish causality in science. > Or perhaps you don’t fully understand the challenges of establishing causality? Just because an intervention causes an improvement in some bio markers that are associated with lower mortality (unfortunately) does not mean that the intervention will cause lower mortality. > The classic example is vitamin D supplements: Higher vitamin D levels are associated with lower mortality in many medical conditions. Vitamin D supplements increase vitamin D levels. But vitamin D supplements seldom lower mortality. That is quite a simplification. The research as far as I know is that too low Vitamin D levels are associated with higher mortality, but that supplementation above a certain level is meaningless (and most caucasians can get those levels through normal sun exposure). So yes vitamin D supplements in general don't improve mortality. Note also the situation here is completely different. The link between mortality and morbidity markers and exercise have been established in other studies. The study here establishes that this directly translates to a correlation between fitness and mortality. So in a way it's the opposite of the vitamin D case. > Why? Probably because vitamin D is produced in the skin when we are in the sun. The more healthy subpopulation of any study will typically spend more time outside, so they will have higher vitamin D levels. But it’s (relative) health that causes higher vitamin D, not the other way around. You realise that you are proclaiming causality here? > The only way to reliably establish causality is really an end-to-end randomized controlled trial (RCT). Stitching together two RCTs is not sufficient. A RCT does not establish causality. That's essentially my point. A single study/experiment never proofs causality. You need a theory to explain the causality and multiple studies that falsify other possible causality mechanisms. That has been done extensively for excercise and morbity/mortality, the current study just establishes that this also correlates in the bigger picture. So yes the study itself does not "proof" causality, it's just a piece in the bigger puzzle of causality. > (Not saying that exercise does not lower mortality BTW, just that it’s complicated, and a study such as this is probably picking up two signals: one causal and one purely correlational.)
indymike 2y ago
> People implying that the study shows only correlation really don't seem to understand how we establish causality in science. This is where explaining how causality is established really helps people understand what they are reading better, especially when confronted with something that is not light reading and summarizes a much larger, even heavier read. In other words, instead of saying people don't understand, maybe help people to understand.

prerok 2y ago

Well, I don't doubt that cardio fitness decreases the chance for morbidity and mortality, but in this study they only established correlation. I think that in science we have to be extra careful just what we are saying. And saying that this study proves causality would be wrong. I am sure there are other studies that could (do?) prove it but this statistical analysis can only prove correlation.

bigfudge 2y ago
We only have correlation for lots of important results. Smoking cigarettes is one example. Experiments are almost always better, and a lot of correlational research is terrible. But knowing something for sure doesn’t always require an experiment.
- caeril 2y ago
  That's a bit different. There's no reasonable path to reverse-causality with smoking. "People prone to developing lung cancer are more likely to take up smoking" doesn't really make sense. On the other hand, "Fat and unhealthy people are more prone not to exercise" is a reasonable reverse-causality position, for all kind of reasons. Your knees hurt, you're self-conscious, it's extremely uncomfortable, your fat jiggles, etc. I'm not saying direct causality doesn't exist here, I'm just saying that reverse-causality is reasonable, whereas with smoking it is not. Assuming causation from this particular correlation is harder than it is with smoking.
  - lukas099 2y ago
    "People who smoke are more likely to do other things that the medical establishment condemns" is a plausible non-causal explanation for the correlation between smoking and lung cancer. I think that we just have to look at the best explanation for the current data to determine causality, but it’s always a guess.
    - dllthomas 2y ago
      That is a plausible explanation for a present correlation, but I think less so for when the correlation was first noted.
tea-coffee 2y ago
From the Methods section of the paper (under "study design"): "We also included meta-analyses that pooled data from primary prospective/retrospective cohort or case-control studies. These studies were the focus because of their ability to assess causality for observational research". While "causality" is a strong word in the sentence above, the data is much stronger than simple correlation. Of course, outright causality has not been established, but the evidence to determine a predictive association is strong.

Fricken 2y ago

Exercise is a four letter word for many on HN. Content posted here touting the various benefits of exercise is always met with mental gymnastics from commenters who would rather do anything but go out and break a sweat.

portaouflop 2y ago
I used to be like that- for some reason I started exercising this year and it completely changed my life. I feel better than ever and it’s not like I’m training a lot, a couple of times per week. If you’re on the fence and reading this definitely give it another shot
- Eddy_Viscosity2 2y ago
  Second this, it doesn't just help your body. Your mental health will also see benefits.
agumonkey 2y ago
Tiny side note in case someone is interested: my health crashed in 2014, I used to be a runner, now I can't jog nor reach sweat levels of efforts. But you don't need to reach that level to restore/improve your health tremendously. - a tiny step to do small one leg squat - a kettle bell or any weight for deadlift doing 10 repetition at low speed, low intensity[0] of the two abose everytime you get bored, anxious, or lost on a youtube rabbithole will feel like nothing, yet, over the week you'll start to feel muscle grow slowly, less joint pain, better posture, better ability to move, everything will feel easier thus having a better mood too. once in the morning, once in the evening.. or maybe more as one sees fit, until you feel the drive to do more (usually a month of slow and pleasurable exercise will naturally lead to a desire to try more) [0] real slow, like taichi slow.. no muscle burn, no fast breathing..
Workaccount2 2y ago
For HN's that want to get active, I cannot suggest getting a (higher tier) Garmin watch enough. I know you have the money so just drop the $800. It's stats on stats on stats and essentially creates a "character health" panel for your own body. Charts, metrics, real time tracking and exportable data. There are achievement badges and social connections if that's your thing too (good for you and your SO). For me, being able to see real-time metrics and real time improvement really amped up my motivation to go out an exercise. Rewards were no longer this mysterious ethereal thing that will maybe show up sometime in the future if I keep grinding. Day by day, I could see improvement and sure enough could feel it too.

huijzer 2y ago

No sorry but this is not how causality works. When event B happens a thousand times after event A, you cannot conclude that A causes B. For example, there could be a confounder, that is, a third event that triggers A and then B (so not A triggers B, but this third event triggers B). See The Book of Why by Pearl for more information on causality.

gjjydfhgd 2y ago
You cannot conclude, but at the same time you should increase your probability of causality. Keeping it unchanged is another kind of logical error.
bigfudge 2y ago
But Pearls book explicitly argues for using correlational evidence to infer causation. The trick is to actually have a causal model and design observational studies that rule out as many alternative explanations as possible.
- cycomanic 2y ago
  Thank you! That seems to be the issue with many commenter here, they believe you can have a single experiment that somehow proofs causality. The same scepticism would not be brought against many "hard science" experiments, even though they do the exact same thing, falsify alternative explanations until they have high certainty that they have causality.
knallfrosch 2y ago
Can you give an example where causality is established then? It sounds like I can hit you a 1,000 times, you feel pain a 1,000 times and you still don't believe there's a causality.
- gwd 2y ago
  Consider this quote: "Every time we send 5 fire trucks to a fire, the damage is 10x than when we only send one fire truck. We've observed this 1000 times. And still you don't believe that the fire trucks are causing the damage." In this case it should be absolutely clear that A (lots of trucks) aren't causing B (lots of damage), but rather a third aspect, C (size of fire) is causing both A and B. Insisting that A causes B will result in completely counterproductive interventions, like "send only one truck to all fires". The same thing could be true for cardiovascular fitness. If people are sick, they're much less likely to running or hiking up a mountain. So rather than poor cardio fitness (A) causing high mortality (B), it could be that a third thing, sickness (C) is causing both A and B. If that is the case, then shaming people who are sick into trying to exercise, instead of making them healthy enough so that they feel like running, is likely to make things worse rather than better. How do you tell the difference? Well the "gold standard" is randomized controlled trials. Pick 3,000 random people. Tell 1000 of them to exercise more, and 1000 of them to exercise less, and 1000 leave alone, and compare. If the "exercise more" group is healthier at the end of 10 years, that's decent evidence that "exercise more" is a useful intervention. Failing that, you can think of other possible confounding factors and control for them. Don't just ask how much they exercise; ask how old they are, and how well they are, how stressed they are, and loads of other factors which might both cause both A and B, and use statistical methods to detect whether one of those factors is actually a better explanation than "A -> B".

bjornsing 2y ago

> People implying that the study shows only correlation really don't seem to understand how we establish causality in science. Or perhaps you don’t fully understand the challenges of establishing causality? Just because an intervention causes an improvement in some bio markers that are associated with lower mortality (unfortunately) does not mean that the intervention will cause lower mortality. The classic example is vitamin D supplements: Higher vitamin D levels are associated with lower mortality in many medical conditions. Vitamin D supplements increase vitamin D levels. But vitamin D supplements seldom lower mortality. Why? Probably because vitamin D is produced in the skin when we are in the sun. The more healthy subpopulation of any study will typically spend more time outside, so they will have higher vitamin D levels. But it’s (relative) health that causes higher vitamin D, not the other way around. The only way to reliably establish causality is really an end-to-end randomized controlled trial (RCT). Stitching together two RCTs is not sufficient. (Not saying that exercise does not lower mortality BTW, just that it’s complicated, and a study such as this is probably picking up two signals: one causal and one purely correlational.)

cycomanic 2y ago
> > People implying that the study shows only correlation really don't seem to understand how we establish causality in science. > Or perhaps you don’t fully understand the challenges of establishing causality? Just because an intervention causes an improvement in some bio markers that are associated with lower mortality (unfortunately) does not mean that the intervention will cause lower mortality. > The classic example is vitamin D supplements: Higher vitamin D levels are associated with lower mortality in many medical conditions. Vitamin D supplements increase vitamin D levels. But vitamin D supplements seldom lower mortality. That is quite a simplification. The research as far as I know is that too low Vitamin D levels are associated with higher mortality, but that supplementation above a certain level is meaningless (and most caucasians can get those levels through normal sun exposure). So yes vitamin D supplements in general don't improve mortality. Note also the situation here is completely different. The link between mortality and morbidity markers and exercise have been established in other studies. The study here establishes that this directly translates to a correlation between fitness and mortality. So in a way it's the opposite of the vitamin D case. > Why? Probably because vitamin D is produced in the skin when we are in the sun. The more healthy subpopulation of any study will typically spend more time outside, so they will have higher vitamin D levels. But it’s (relative) health that causes higher vitamin D, not the other way around. You realise that you are proclaiming causality here? > The only way to reliably establish causality is really an end-to-end randomized controlled trial (RCT). Stitching together two RCTs is not sufficient. A RCT does not establish causality. That's essentially my point. A single study/experiment never proofs causality. You need a theory to explain the causality and multiple studies that falsify other possible causality mechanisms. That has been done extensively for excercise and morbity/mortality, the current study just establishes that this also correlates in the bigger picture. So yes the study itself does not "proof" causality, it's just a piece in the bigger puzzle of causality. > (Not saying that exercise does not lower mortality BTW, just that it’s complicated, and a study such as this is probably picking up two signals: one causal and one purely correlational.)

indymike 2y ago

> People implying that the study shows only correlation really don't seem to understand how we establish causality in science. This is where explaining how causality is established really helps people understand what they are reading better, especially when confronted with something that is not light reading and summarizes a much larger, even heavier read. In other words, instead of saying people don't understand, maybe help people to understand.