4 ms·

Please find the (apparently obvious) problem with the results. I keep seeing articles about how study after study confirms the alcohol/longevity link, so if yo

Please find the (apparently obvious) problem with the results. I keep seeing articles about how study after study confirms the alcohol/longevity link, so if you can find the common problem those studies all have, (very minor) fame awaits! :)

fjh 16y ago
The one problem that I can see is the fact that these studies can only correlate people's behaviour and their longevity, as opposed to randomly assigning people into groups of drinkers/teetotallers. It is a fair assumption that people who don't drink differ from other people in other aspects, too. (e.g. they might be more religious, or not drink for health reasons, or have a less active social life etc). You obviously cannot do double-blind studies about long-term alcohol consumption, so correlation is the best we have, but one must still be careful when interpreting these kinds of results.
- ojbyrne 16y ago
  There are numerous experimental designs that, while they don't have the predictive value of double-blind experiments, are found to have significant predictive value. The simplest way is to control for the confounding factors such as you describe. Google "Quasi-experimentation."
- jmatt 16y ago
  Full disclojure (hah): I don't drink though I don't consider myself a teetotaler. I really don't mind what others do to their bodies. not drink for health reasons Yes this sounds plausible. Even if the person isn't unhealthy maybe their family history is such that they still don't drink. have a less active social life etc As a semi-introvert[1] I always wonder how true this is for the minority of us that much prefer this so called social isolation. I'd take a night alone or with a small group of friends over a bar or conference absolutely every time always. Of course that is assuming there are no other external influences - which is rarely (I'd say never but we're on HN...) the case. All that really matters is whether you are suffering from anxiety rather than loneliness[2]. [1] I'm XNTP somewhere between INTP and ENTP on the meyer brigs http://en.wikipedia.org/wiki/Myers-Briggs_Type_Indicator http://en.wikipedia.org/wiki/Myers-Briggs_Type_Indicator [2] I can't remember the last time I was lonely or bored. Hmmm, is this really that unusual amongst similar introverts? I just consider myself somewhat curious and self-motivated.
Evgeny 16y ago
Those who have already died from alcohol related causes can not be included into the study, which would much reduce the longevity among alcohol users?
- chc 16y ago
  It also doesn't account for people who previously died from lack of alcohol. I doubt either group is worth considering.
  - Evgeny 16y ago
    I think the argument may exist that if you didn't die from alcohol abuse early enough, then your body might have higher than average resistance towards various harmful substances. Therefore, alcohol usage is just correlation, while the real reason for increased longevity is the said resistance.
    - chc 16y ago
      This would require evidence of many young deaths from alcohol abuse. As far as I know, even heavy drinkers are pretty unlikely to die from their habit before age 50. Moderate drinkers, reflected as the healthiest in the study, are tremendously unlikely to die from alcohol abuse early in their lives. And again, maybe the reason non-drinkers show as much longevity as they do is because those are the ones who have an increased resistance to the negative effects of abstinence. Maybe most of them die much younger. Since non-drinkers are such a minority in most places, this is actually more likely, since the die-off would be less noticeable than if everybody who drank an occasional beer dropped dead at age 40. (I still don't think this is likely either, incidentally.)
      - silentbicycle 16y ago
        > This would require evidence of many young deaths from alcohol abuse. As far as I know, even heavy drinkers are pretty unlikely to die from their habit before age 50. Don't forget people who have died as a result of accidents while intoxicated (driving drunk, etc.) or have been incarcerated for drunken infractions. They may be excluded from those statistics, adding a kind of bias. Not sure if the total is high enough to impact the study's results, but it's worth checking.

fjh 16y ago

The one problem that I can see is the fact that these studies can only correlate people's behaviour and their longevity, as opposed to randomly assigning people into groups of drinkers/teetotallers. It is a fair assumption that people who don't drink differ from other people in other aspects, too. (e.g. they might be more religious, or not drink for health reasons, or have a less active social life etc). You obviously cannot do double-blind studies about long-term alcohol consumption, so correlation is the best we have, but one must still be careful when interpreting these kinds of results.

ojbyrne 16y ago
There are numerous experimental designs that, while they don't have the predictive value of double-blind experiments, are found to have significant predictive value. The simplest way is to control for the confounding factors such as you describe. Google "Quasi-experimentation."
jmatt 16y ago
Full disclojure (hah): I don't drink though I don't consider myself a teetotaler. I really don't mind what others do to their bodies. not drink for health reasons Yes this sounds plausible. Even if the person isn't unhealthy maybe their family history is such that they still don't drink. have a less active social life etc As a semi-introvert[1] I always wonder how true this is for the minority of us that much prefer this so called social isolation. I'd take a night alone or with a small group of friends over a bar or conference absolutely every time always. Of course that is assuming there are no other external influences - which is rarely (I'd say never but we're on HN...) the case. All that really matters is whether you are suffering from anxiety rather than loneliness[2]. [1] I'm XNTP somewhere between INTP and ENTP on the meyer brigs http://en.wikipedia.org/wiki/Myers-Briggs_Type_Indicator http://en.wikipedia.org/wiki/Myers-Briggs_Type_Indicator [2] I can't remember the last time I was lonely or bored. Hmmm, is this really that unusual amongst similar introverts? I just consider myself somewhat curious and self-motivated.

Evgeny 16y ago

Those who have already died from alcohol related causes can not be included into the study, which would much reduce the longevity among alcohol users?

chc 16y ago
It also doesn't account for people who previously died from lack of alcohol. I doubt either group is worth considering.
- Evgeny 16y ago
  I think the argument may exist that if you didn't die from alcohol abuse early enough, then your body might have higher than average resistance towards various harmful substances. Therefore, alcohol usage is just correlation, while the real reason for increased longevity is the said resistance.
  - chc 16y ago
    This would require evidence of many young deaths from alcohol abuse. As far as I know, even heavy drinkers are pretty unlikely to die from their habit before age 50. Moderate drinkers, reflected as the healthiest in the study, are tremendously unlikely to die from alcohol abuse early in their lives. And again, maybe the reason non-drinkers show as much longevity as they do is because those are the ones who have an increased resistance to the negative effects of abstinence. Maybe most of them die much younger. Since non-drinkers are such a minority in most places, this is actually more likely, since the die-off would be less noticeable than if everybody who drank an occasional beer dropped dead at age 40. (I still don't think this is likely either, incidentally.)
    - silentbicycle 16y ago
      > This would require evidence of many young deaths from alcohol abuse. As far as I know, even heavy drinkers are pretty unlikely to die from their habit before age 50. Don't forget people who have died as a result of accidents while intoxicated (driving drunk, etc.) or have been incarcerated for drunken infractions. They may be excluded from those statistics, adding a kind of bias. Not sure if the total is high enough to impact the study's results, but it's worth checking.